Index of sw-kernelinferencelibrary-public-generic-local/792869289f5e46f83b26f85e550c322e2ea8a9d7/fmha/trtllm-gen
Name Last modified Size
../
include/ 08-Jun-2026 17:43 -
checksums.txt 08-Jun-2026 17:43 2.93 MB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.17 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 77.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 72.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 83.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 78.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.92 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.31 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.92 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 81.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.06 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.01 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 59.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 79.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 72.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.06 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 81.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 77.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 72.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 81.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.78 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.17 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 72.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.31 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 79.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.53 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.92 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.06 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.17 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.40 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 72.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 78.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 72.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.40 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 80.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.40 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.06 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.28 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.53 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 72.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.28 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.22 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.03 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.28 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.01 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 84.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 77.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 90.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 83.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.47 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.31 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.92 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 82.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 87.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 81.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.22 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 80.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.01 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 64.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 86.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 80.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.81 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.01 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 83.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 89.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 82.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.47 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 93.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 86.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 84.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 98.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 85.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.56 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.22 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.56 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.56 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 98.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.53 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.47 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.22 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.53 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 91.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 89.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.22 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 91.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 182.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 179.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 148.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 181.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 183.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 179.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 151.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 172.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 168.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 137.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 63.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 54.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 59.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.01 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 61.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 52.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 65.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 56.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 72.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.50 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 58.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 60.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 51.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 56.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 48.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 63.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 55.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.26 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 69.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 61.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 61.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 53.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 65.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 58.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.98 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.47 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 69.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.20 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.67 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 66.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 59.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.40 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 65.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 58.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 56.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 49.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.17 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 64.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.42 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 72.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.73 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.81 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.97 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.90 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.56 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.12 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 56.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.62 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.31 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.60 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.87 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.34 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 55.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 49.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.48 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.35 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.15 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.65 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.59 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.23 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.27 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.93 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.63 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 77.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.10 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.46 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 59.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.30 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 82.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 73.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.70 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.76 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.49 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.09 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.43 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.00 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.68 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.83 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.72 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 57.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.05 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.11 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 79.99 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.94 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 72.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 61.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.45 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.36 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.89 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.29 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.18 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.88 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.13 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.19 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.75 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.58 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.47 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 66.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 56.14 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.86 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 78.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.57 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.69 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.41 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.39 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.66 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.24 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.95 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.71 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.33 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.64 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.74 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.25 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.28 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.04 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.81 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 72.54 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.51 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.84 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.91 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.55 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.82 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.38 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.21 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.85 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.08 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 58.77 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.80 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.61 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 81.37 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.32 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.44 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.52 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.96 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.16 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.07 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.02 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.79 KB
fmhaSm100aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.17 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.92 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 88.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 89.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 88.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.33 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 80.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.31 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.72 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.23 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.36 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.43 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 83.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 73.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 86.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 76.35 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 80.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.23 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.98 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.23 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.98 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.20 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.20 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.20 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 78.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.36 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.98 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.98 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.98 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.23 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.36 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.72 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.02 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.02 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.02 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.97 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 94.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.81 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.64 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.81 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 76.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.33 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.81 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.23 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.11 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.02 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.77 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 83.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 73.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.26 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 68.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.43 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 64.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.36 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.33 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 66.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 69.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.43 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 64.58 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.51 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 86.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 80.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.63 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.85 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.93 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.08 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 80.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 73.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 56.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.36 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.88 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.76 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.64 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 89.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.73 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.64 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.83 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.14 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.99 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.49 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 98.61 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.67 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 104.99 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 100.55 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.47 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 102.42 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 121.25 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 108.51 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.36 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 104.04 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 98.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 104.99 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 100.55 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.59 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 102.55 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 121.38 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 108.51 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.48 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 104.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.08 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 96.28 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.50 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 99.85 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.05 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 98.15 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.05 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 100.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 123.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 103.75 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.40 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 101.84 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.20 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 99.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.17 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 98.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.18 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 100.22 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 123.58 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 103.88 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.52 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 101.96 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.82 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.60 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.21 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.82 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.60 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.00 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.53 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.01 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.69 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.11 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.82 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.17 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.93 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.78 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.36 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.32 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.53 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.90 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.96 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.92 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.36 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.11 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.92 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.83 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.36 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.93 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.46 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.82 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 139.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.78 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.69 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.89 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.14 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.17 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.83 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.92 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.90 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.85 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 225.89 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 219.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 222.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 217.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 189.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 205.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 180.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 200.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 209.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 203.36 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 205.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 201.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 172.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 190.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.35 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 202.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 195.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 198.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 193.93 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 167.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 183.15 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 159.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 230.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 223.69 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 226.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 221.64 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 192.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.92 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 185.32 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 213.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 207.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 210.17 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 205.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 175.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 169.46 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.89 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 206.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 200.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 202.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 197.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 170.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.17 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.61 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.14 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.46 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.69 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.15 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.82 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.32 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.01 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.49 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.43 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.92 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.35 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.35 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.36 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.54 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.64 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.61 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.96 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.64 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.32 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.01 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.60 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.78 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.69 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.89 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.14 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.23 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.12 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.94 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 128.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 135.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 143.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 141.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 127.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 148.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 127.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 131.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 143.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 142.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 123.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 128.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 96.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 145.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 135.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 143.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 123.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 134.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 130.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 128.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 132.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 103.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 151.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 141.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 132.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 145.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 160.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 135.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 144.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 135.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 138.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 138.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 145.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 129.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 164.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 152.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 153.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 142.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 156.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 142.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 153.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 140.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 133.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 149.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 165.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 131.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 155.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 139.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 136.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 141.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 142.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 86.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 141.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 137.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 153.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 159.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 150.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 150.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 153.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 145.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 162.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 153.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 152.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 151.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 131.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 130.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 135.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 131.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 184.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 193.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 179.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 170.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 182.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 182.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 190.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 161.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 170.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 173.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 167.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 149.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 155.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 158.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 149.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 161.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 152.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 138.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 141.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 161.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 159.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 159.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 157.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 159.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 151.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 158.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 150.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 156.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 148.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 156.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 135.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 136.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 231.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 222.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 226.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 240.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 219.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 201.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 191.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 223.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 216.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 190.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 184.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 159.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 192.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 183.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 188.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 203.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 181.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 176.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 200.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 188.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 163.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 147.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 177.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 168.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 173.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 188.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 162.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 184.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 149.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 210.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 203.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 211.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 316.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 327.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 311.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 184.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 170.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 195.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 182.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 191.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 182.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 173.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 235.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 226.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 231.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 245.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 223.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 206.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 196.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 228.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 220.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 195.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 188.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 197.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 187.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 192.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 208.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 181.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 205.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 192.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 170.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 132.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 153.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 182.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 172.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 177.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 192.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 166.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 188.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 174.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 155.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 99.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 98.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 97.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 238.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 230.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 229.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 215.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 87.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 75.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 142.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 127.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 160.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 174.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 160.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 146.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 145.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 159.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 171.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 137.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 151.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 141.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 132.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 143.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 128.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 144.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 171.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 155.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 171.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 158.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 175.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 144.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 157.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 143.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 155.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 162.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 174.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 144.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 152.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 154.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 141.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 136.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 145.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 157.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 135.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 138.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 143.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 148.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 196.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 177.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 184.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 194.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 160.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 177.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 173.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 151.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 187.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 181.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 163.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 142.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 167.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 142.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 139.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 167.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 148.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 142.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 145.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 144.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 206.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 225.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 200.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 149.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 164.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 153.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 164.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 204.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 220.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 198.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 194.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 191.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 163.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 211.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 208.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 179.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 159.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 162.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 143.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 157.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 187.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 184.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 157.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 128.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 164.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 158.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 147.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 158.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 174.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 195.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 168.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 132.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 136.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 172.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 188.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 169.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 149.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 186.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 142.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 137.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 127.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 162.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 142.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 146.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 138.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 130.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 180.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 125.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 129.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 173.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 171.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 158.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 134.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 119.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 148.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 138.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 123.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 145.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 162.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 169.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 160.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 139.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 146.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 246.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 188.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 238.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 180.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 196.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 197.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 243.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 266.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 207.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 236.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 178.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 156.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 229.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 230.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 187.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 173.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 257.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 253.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 210.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 195.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 195.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 191.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 220.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 222.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 178.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 159.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 142.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 183.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 191.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 138.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 203.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 195.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 155.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 158.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 200.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 222.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 189.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 192.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 198.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 187.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 169.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 222.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 210.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 191.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 160.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 149.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 188.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 179.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 160.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 145.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 158.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 152.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 185.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 177.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 141.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 182.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 205.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 181.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 174.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 179.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 161.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 203.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 192.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 183.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 141.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 170.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 152.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 144.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 175.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 168.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 176.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 181.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 175.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 188.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 183.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 170.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 156.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 94.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 100.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 96.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 103.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 110.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 98.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 104.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 100.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 78.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 130.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 215.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 204.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 204.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 190.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 89.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 141.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 127.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 94.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 100.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 110.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 98.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 104.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 100.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 138.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 97.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 94.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 97.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 96.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 147.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 134.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 140.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 123.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 132.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 97.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta128PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 130.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 118.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 119.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 118.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 104.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 137.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 138.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 128.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 161.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 166.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 179.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 170.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 190.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 155.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 141.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 154.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 155.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 174.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 113.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 168.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 156.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 145.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 155.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 137.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvDenseDynamicTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 174.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 183.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 175.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 194.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 179.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 146.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 155.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 159.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 179.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 162.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 152.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 173.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 156.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 142.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 87.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 93.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 89.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 91.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 97.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 93.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 86.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 94.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 96.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 209.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 201.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 200.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 185.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 85.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 74.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 83.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 110.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 87.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 112.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 93.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 89.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 91.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 97.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 93.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 103.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 141.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 127.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 89.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 91.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 142.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 148.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 142.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 132.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 135.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 134.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 131.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 96.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 99.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 125.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 134.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 138.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 151.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 137.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 132.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 132.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 89.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 153.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 108.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 111.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 164.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 143.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 136.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 167.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 149.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 116.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 103.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 143.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 119.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 105.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 157.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 148.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 101.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 104.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 136.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 153.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 139.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 144.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 108.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 95.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 111.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 152.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 96.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 99.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 147.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 141.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 96.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 90.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 105.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:44 93.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 106.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 117.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 08-Jun-2026 17:43 100.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 175.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 167.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 144.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 172.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 194.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 170.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 164.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 174.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 136.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 198.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 182.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 173.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 134.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 165.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 141.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 137.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 158.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 177.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 156.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 143.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 179.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 164.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 166.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 146.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 142.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 133.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 147.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 159.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 148.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 134.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 172.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 162.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 139.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 144.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 140.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 150.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 133.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 127.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 101.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 109.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 96.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 102.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 98.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 99.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 106.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 101.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 126.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 124.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 137.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 99.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 136.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 140.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 143.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 121.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 124.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 134.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 88.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 131.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 94.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 86.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 127.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 117.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 85.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 130.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 121.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 132.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 96.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 102.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 98.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 100.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 119.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 106.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 115.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 101.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 97.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 137.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 139.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 126.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 128.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 98.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 109.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 101.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 120.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 130.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 111.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 142.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 139.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 122.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 119.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 134.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 144.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 131.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 100.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 102.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 136.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 106.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 106.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 104.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 107.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 135.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 133.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 130.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 125.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 106.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 107.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 131.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 131.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 128.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 117.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 140.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 126.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 132.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 138.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 130.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 139.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 133.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 119.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 129.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 125.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 171.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 182.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 179.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 194.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 180.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 160.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 129.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 143.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 115.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 156.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 147.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 115.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 152.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 145.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 164.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 179.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 163.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 145.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 132.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 122.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 136.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 108.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 141.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 112.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 105.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 138.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 157.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 142.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 172.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 157.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 138.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 105.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 176.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 167.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 172.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 187.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 165.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 183.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 169.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 199.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 185.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 169.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 133.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 128.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 160.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 151.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 157.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 172.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 149.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 127.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 168.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 183.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 168.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 154.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 136.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 126.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 111.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 141.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 122.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 150.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 124.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 161.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 146.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 177.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 147.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 129.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 124.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 108.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 119.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 112.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 86.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 100.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 89.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 106.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 95.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 102.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 90.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 104.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 92.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 98.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 106.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 94.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 112.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 114.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 111.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 96.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 100.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 108.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 108.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 113.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 102.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 109.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 114.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 101.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 114.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 120.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 125.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 99.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 124.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 109.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 104.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 91.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 110.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 126.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 109.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 120.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 118.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 116.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 104.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 123.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 103.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 94.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 129.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 105.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 119.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 107.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 107.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 98.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 120.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 103.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 114.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 127.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 116.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 101.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 105.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 115.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 117.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 116.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 121.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 100.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 89.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 106.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 95.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 102.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 91.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 104.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 92.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 99.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 106.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 94.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 120.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 108.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 126.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 114.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 122.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 124.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 111.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 130.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 118.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 126.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 114.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 118.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 106.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 121.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 116.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 107.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 128.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 95.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 110.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 103.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 100.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 110.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 92.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 82.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 95.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 104.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 97.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 103.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 138.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 122.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 112.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 101.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 131.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 100.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 93.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 134.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 113.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 102.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 123.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 93.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 115.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 123.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 111.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 124.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 110.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 137.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 125.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 113.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 102.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 98.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 134.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 120.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 120.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 108.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 126.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 114.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 122.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 110.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 124.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 112.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 130.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 118.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 126.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 114.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 101.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 92.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 102.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 143.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 93.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 148.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 150.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 138.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 140.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 129.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 125.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 144.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 119.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 128.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 144.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 130.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 140.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 135.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 126.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 143.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 136.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 129.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 141.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 127.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 144.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 130.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 119.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 142.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 143.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 134.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 129.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 170.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 179.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 167.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 184.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 169.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 155.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 164.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 152.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 168.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 144.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 158.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 145.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 162.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 175.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 184.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 171.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 191.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 174.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 163.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 138.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 169.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 156.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 176.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 158.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 148.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 152.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 141.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 169.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 142.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 128.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 86.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 104.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 92.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 98.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 94.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 107.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 96.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 114.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 102.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 98.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 86.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 93.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 78.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 104.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 92.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 98.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 106.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 94.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 108.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 96.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 114.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 102.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 104.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 92.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 107.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 94.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 109.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 96.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 103.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 110.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 98.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 117.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 104.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 100.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 92.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 117.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 112.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 120.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 117.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 107.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 94.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 109.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 96.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 103.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 98.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 117.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 104.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 100.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 92.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 96.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 85.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 99.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 87.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 97.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 85.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 124.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 138.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 128.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 134.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 143.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 136.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 138.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 131.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 124.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 122.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 125.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 137.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 138.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 129.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 170.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 179.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 171.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 191.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 176.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 155.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 142.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 164.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 156.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 175.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 160.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 144.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 158.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 169.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 174.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 184.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 175.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 195.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 179.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 164.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 147.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 147.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 169.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 148.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 160.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 180.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 162.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 153.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 141.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 153.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 174.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 156.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 142.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 137.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 95.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 83.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 99.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 87.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 95.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 87.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 93.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 89.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 83.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 99.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 91.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 117.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 97.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 93.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 87.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 98.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 95.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 94.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 96.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 107.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 95.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 99.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 87.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 93.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 113.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 89.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:44 111.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:44 91.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:44 117.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:44 97.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:44 93.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.46 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.03 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.39 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 149.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 136.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 135.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.39 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.06 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.22 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 144.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.03 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.07 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 127.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.03 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 129.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.42 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.79 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.99 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.47 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.06 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.37 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.42 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.00 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.07 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 150.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 130.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.21 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.21 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.99 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.96 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.06 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.46 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.60 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 144.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.60 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.00 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.39 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 146.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 154.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 134.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.96 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 134.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 127.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 127.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.47 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.42 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 144.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 141.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 132.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.00 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.00 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 150.21 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.47 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 128.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 131.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 138.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 135.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 127.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 146.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.14 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.37 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.06 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 140.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 149.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.42 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 131.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.92 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 120.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 128.46 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 134.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 125.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 94.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 111.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 102.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 87.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 91.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 102.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 105.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 93.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 147.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 152.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 147.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 158.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 142.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.06 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.68 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 142.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 133.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.10 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.68 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 145.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 161.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 134.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 145.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 136.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 126.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 124.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 147.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.06 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.82 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 142.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.10 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.72 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 107.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 147.06 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.67 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.67 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.68 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.67 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.92 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.43 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.21 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.50 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.21 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 143.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 128.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 146.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 131.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.82 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 119.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 155.00 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.21 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 143.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 147.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 157.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.10 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 144.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.14 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 154.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 141.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 134.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.04 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 134.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 166.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.94 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 156.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 140.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 137.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.10 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.18 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 130.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 152.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 142.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 137.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 151.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 136.48 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 127.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 138.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.82 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 121.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 136.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 130.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.08 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.94 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 143.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 131.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 141.72 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.68 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 124.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.40 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 144.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 118.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.93 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 117.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.10 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 110.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 98.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.89 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.75 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 154.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 155.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 155.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 153.85 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 160.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 151.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 151.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.89 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 154.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 147.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 163.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 154.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 154.07 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.25 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 152.33 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 141.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 137.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.36 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 124.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 131.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.42 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 140.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 136.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 127.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 132.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.04 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 124.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.77 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 159.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 150.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 162.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 153.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 129.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 142.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 132.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 119.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 162.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 161.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 160.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 159.24 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 160.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 153.26 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 159.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 151.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 157.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 149.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 157.48 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 145.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 139.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 140.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 136.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 140.96 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 137.85 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.00 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.00 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 130.15 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.01 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 126.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.97 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 231.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 222.33 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 141.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 134.99 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 226.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 241.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 219.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 202.04 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 192.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 224.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 216.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 190.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 184.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 140.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 159.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.65 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 129.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 192.86 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 183.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 188.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 204.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 180.99 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.13 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 135.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 176.65 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 167.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 201.89 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 188.76 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 163.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 127.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 148.87 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.56 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 116.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 177.77 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.11 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 173.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 188.89 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.90 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 162.83 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 150.33 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 185.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 141.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.93 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 109.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.08 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 235.86 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 226.48 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 138.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 231.25 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 245.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 223.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 137.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 206.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 196.48 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 229.96 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 220.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 195.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 188.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 143.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 165.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 145.75 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 133.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 197.26 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 187.76 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 133.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.51 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 192.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 208.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 185.15 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 140.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 181.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 206.15 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 192.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 170.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 158.75 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.82 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 154.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.60 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 182.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 172.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 177.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 193.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 170.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 134.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 166.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 154.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 190.08 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 175.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 155.42 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 141.99 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 125.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 146.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 102.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 101.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 90.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 109.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 120.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 119.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 126.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 121.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.89 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.15 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.15 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.06 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 97.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.60 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 101.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 115.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.64 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.61 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 114.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 131.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 122.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.18 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.64 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.89 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 114.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 99.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 133.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.08 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 123.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 116.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 106.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.15 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 114.06 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 103.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 122.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 111.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 130.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.06 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.90 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 138.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 121.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.64 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.01 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 120.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 117.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.37 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 123.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 137.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 127.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 124.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.15 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 115.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 135.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.61 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.08 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 110.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 117.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 115.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 120.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 134.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 124.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.18 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 134.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.11 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 116.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.61 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 115.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 114.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 132.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 123.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 120.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 111.18 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 129.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 137.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 109.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 104.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 128.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 118.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 116.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 121.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.53 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.15 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 134.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 125.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 113.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.69 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.00 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.27 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 112.42 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 97.13 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 106.28 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.02 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 113.62 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 96.83 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 116.66 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 100.20 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 123.12 KB
fmhaSm100fKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 108.03 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 127.68 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.94 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 134.23 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.86 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.89 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 105.55 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.43 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.60 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.65 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 102.05 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.06 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.98 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 128.20 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 110.66 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 135.63 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 118.59 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 121.69 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 104.88 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 129.00 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 112.82 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 118.58 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 101.92 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 125.88 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 109.86 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.23 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 122.67 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 145.67 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 129.28 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 133.43 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 117.29 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 139.87 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 123.91 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 130.20 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 113.66 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 136.62 KB
fmhaSm100fKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 120.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.92 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 78.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 69.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 63.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 83.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.26 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 78.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 70.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.17 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.67 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 75.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 67.53 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.51 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 70.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 61.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 81.67 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 73.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.01 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 77.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 68.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.92 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.65 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 76.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 74.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 66.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.26 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 69.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 60.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 55.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 80.42 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 75.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 67.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.17 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.17 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 76.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 68.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 71.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 62.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 57.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 82.17 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 77.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 69.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.65 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.31 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 76.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 71.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 81.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 73.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 69.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 80.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 78.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.06 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 68.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 78.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.81 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 103.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.53 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.31 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 74.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 70.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 67.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 80.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 74.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.81 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.81 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 68.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.53 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 61.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.56 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 68.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.65 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.26 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.81 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.56 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.65 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 107.67 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 85.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 76.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 78.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 69.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 90.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 84.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 96.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 105.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 97.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 82.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 74.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 75.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 66.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 88.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 81.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 73.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.53 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 81.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 74.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 65.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 68.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 87.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 78.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 80.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 74.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 108.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 102.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 84.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 77.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 68.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 90.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 83.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 127.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 125.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 103.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 94.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.42 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 111.51 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 107.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 117.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 108.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.42 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 119.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 86.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 88.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 78.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 105.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 102.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.31 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 93.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.92 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 117.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 84.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 85.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 76.17 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 103.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 88.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.92 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 114.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 86.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.51 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 110.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 100.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 114.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 116.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 98.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 92.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 95.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 107.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 112.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 98.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 132.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 119.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 132.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 128.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 118.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 97.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 108.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 99.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 91.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 119.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 111.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 96.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 122.56 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 108.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 113.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 106.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.42 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 112.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 90.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 101.56 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 93.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 85.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 105.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 107.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 94.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 103.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 104.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 123.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 110.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 98.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 124.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 120.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 118.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 109.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 88.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 99.06 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 88.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 90.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 110.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 99.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 102.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 113.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 100.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 104.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 92.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 97.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 105.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 106.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 117.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 126.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 126.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 113.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 95.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 100.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 127.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 115.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 122.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 120.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 111.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 89.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 100.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 90.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 92.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 112.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 101.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 104.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 91.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 115.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 102.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 106.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 95.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 96.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 184.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 179.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 181.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 178.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 149.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 167.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 145.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 163.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 136.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 155.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 130.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 151.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 165.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 161.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 162.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 159.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 132.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 149.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 126.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 146.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 187.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 182.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 184.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 181.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 154.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 143.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 148.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 139.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 174.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 169.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 171.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 139.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 129.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 168.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 164.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 166.26 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 162.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 135.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 125.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 130.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 121.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 80.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 76.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 70.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 63.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 55.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 74.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 64.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 55.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 52.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 75.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 66.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 67.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 67.86 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 59.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 57.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 68.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 73.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 69.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 60.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 61.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 53.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 61.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 53.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 58.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 50.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 73.67 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 65.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 75.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 65.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 57.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 55.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 71.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 76.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 73.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 67.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 78.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 68.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 59.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 52.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 60.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 51.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 57.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 49.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.42 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 64.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 74.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 64.36 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 56.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 61.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.92 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.16 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 70.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 61.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 72.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 61.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 62.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 54.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 59.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 74.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 65.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 66.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 58.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.59 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 70.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 55.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.01 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.46 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 67.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 81.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 87.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.72 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.69 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 82.53 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 67.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 67.76 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 58.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 51.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.75 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 71.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 85.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.44 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 66.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 59.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 58.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 51.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.51 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 56.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 49.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 70.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 64.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 57.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 55.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 79.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 79.31 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 83.90 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 68.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 61.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 68.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 66.51 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 72.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 57.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 54.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 53.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.15 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.88 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.65 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 56.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 50.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.84 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 51.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 55.71 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 49.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 61.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 54.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.01 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 64.56 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.78 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.40 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 57.96 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 51.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDensePagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 79.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.48 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 94.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 77.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 84.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.62 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 83.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.08 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 75.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 89.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 79.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 70.29 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.63 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 60.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 65.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 83.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 74.10 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 73.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 64.18 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 69.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 60.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.28 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 77.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.11 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 91.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.02 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 82.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 75.83 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 66.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.12 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 81.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 76.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 67.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.38 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 57.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 53.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 80.77 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.23 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 61.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 80.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.07 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 72.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.05 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.85 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.03 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 76.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 65.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 90.54 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.89 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 81.37 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 80.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.82 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.58 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 72.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.87 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.20 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.09 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 77.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 75.60 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 66.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.24 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.13 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 56.68 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.33 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.39 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:44 79.52 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 70.98 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.22 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 78.99 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.94 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 60.80 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.64 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.25 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.79 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.66 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.06 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 78.34 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 72.95 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.04 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.00 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.45 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.70 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.91 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 78.35 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.41 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 66.74 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.43 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 77.50 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.19 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 67.61 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 59.30 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.21 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqKOnlyTransQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.14 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 82.27 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 73.73 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 70.97 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 62.93 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 81.49 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.57 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.47 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ64Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.55 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 67.32 KB
fmhaSm103aKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSeparateTransKVQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 59.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.01 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.76 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 89.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 81.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 93.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 91.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 89.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 92.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.33 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 97.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 93.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 80.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 94.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 81.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 98.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 88.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 95.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 102.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 88.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 84.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 76.55 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 83.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 84.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 77.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 82.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 83.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 76.30 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 83.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 89.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 94.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 80.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 97.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 89.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 77.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 74.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 79.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 71.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 93.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 67.30 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 72.51 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 63.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 91.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 70.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 84.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 74.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 69.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 53.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 93.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 95.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 86.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 76.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 87.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 73.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 75.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 68.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 89.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 94.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 92.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 59.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 67.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 93.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.55 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 63.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 54.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 68.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.44 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 56.14 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 72.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.86 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 62.94 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.14 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 101.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 58.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 64.19 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.30 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:44 82.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 98.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 93.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 108.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 87.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 79.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 91.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 87.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:44 100.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 81.62 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 73.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 86.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 83.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 85.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 82.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 77.62 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 75.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 89.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 85.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 79.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 76.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 101.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 97.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 99.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 96.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 94.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 83.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 90.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 81.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 94.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 92.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 89.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 87.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 76.30 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 84.19 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 86.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 84.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 83.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 80.19 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 92.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 88.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 90.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 87.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 82.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.55 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 79.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.14 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 84.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 76.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.30 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 83.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 90.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 78.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 92.53 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 71.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 77.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 69.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 74.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 67.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.12 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 71.01 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.96 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 76.96 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 69.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 85.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.03 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.96 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:44 62.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.62 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.51 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 72.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 84.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.89 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.28 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 78.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 68.12 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 61.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:44 60.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 65.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 58.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 58.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 51.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 70.03 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 63.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 86.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 75.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 88.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 80.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedMultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 82.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 62.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.12 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.01 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ32Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 60.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.96 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 08-Jun-2026 17:43 71.78 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCustomPagedVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 08-Jun-2026 17:43 65.11 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 87.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 82.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 57.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 84.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 76.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 63.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 64.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:44 62.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 81.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 74.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.37 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 57.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 51.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.86 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 83.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 75.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 61.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 55.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 59.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDensePagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 53.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 74.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 72.49 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 73.70 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 94.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 70.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 59.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 67.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 57.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 91.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 66.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.39 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 66.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 67.57 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 65.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 88.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 81.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 63.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 60.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 52.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 71.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 69.36 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 70.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedMultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 68.54 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 90.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.19 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 64.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ16Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 56.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128PersistentSwapsAbForGen.cubin 08-Jun-2026 17:43 62.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ8Kv128StaticSwapsAbForGen.cubin 08-Jun-2026 17:43 54.52 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.48 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.98 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.47 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.25 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.55 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.08 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.24 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.93 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.93 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.19 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.75 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.22 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.67 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.49 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.71 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.49 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.49 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.39 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.37 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.19 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.93 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.25 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.74 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.91 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.87 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.76 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.34 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.61 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 141.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 144.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.34 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.39 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.14 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.15 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 144.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 129.94 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.69 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.04 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.86 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.24 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.94 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 152.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 133.53 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 156.06 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 136.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.68 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.77 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.94 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 160.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 150.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 163.48 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 153.89 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 140.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 127.58 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.76 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.06 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 159.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 153.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 162.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 156.37 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.74 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.69 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.08 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.48 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.74 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.07 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.34 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.47 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.29 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.63 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.79 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.74 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.74 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.17 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.32 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.39 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.04 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.51 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.58 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.86 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.89 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.55 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.68 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.47 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.08 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.14 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.15 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.43 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.75 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.48 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.24 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.43 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.88 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.32 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.58 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.94 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.86 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.24 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 127.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.91 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.86 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.37 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.31 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.69 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.04 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.43 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.55 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.89 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 107.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 103.03 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.32 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 104.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 122.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 111.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 117.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 106.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.26 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.60 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.26 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.11 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.13 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.81 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.81 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.51 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.03 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.42 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.28 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.68 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 107.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 104.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 122.25 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 111.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 106.92 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.51 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.36 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.94 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.61 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.61 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.92 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.21 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.83 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.89 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.81 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.01 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.25 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.54 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.96 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.67 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.68 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.17 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.69 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.28 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.17 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 140.11 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.25 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.08 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.72 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.72 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.51 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 94.89 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 90.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 92.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 98.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 94.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.13 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.42 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.17 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.89 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.72 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.68 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.78 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.00 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.84 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.32 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.50 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.74 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.31 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 95.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.74 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.28 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 92.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.14 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 98.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 94.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk128HV64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.96 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.60 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.32 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.90 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.61 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.19 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.50 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.46 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 102.72 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 120.65 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 109.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 104.78 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 106.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 124.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 112.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 120.19 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 108.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.71 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.46 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.57 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.03 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.15 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.94 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.63 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.99 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.21 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.18 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 102.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 120.65 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 109.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 104.78 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 106.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 124.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 112.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 120.19 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 108.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.71 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.71 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.74 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 98.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 86.18 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.72 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.00 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.67 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.43 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.32 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.92 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.29 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.96 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.83 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.46 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.72 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.73 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.15 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.07 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.74 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.74 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.50 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.49 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.57 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.18 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 141.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.31 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.31 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.15 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.97 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 103.69 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.72 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 101.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 90.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 96.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 103.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 92.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 105.63 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 94.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.64 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.64 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.56 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.26 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.65 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.09 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.73 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.09 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.99 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.56 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.42 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.82 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.07 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.28 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.09 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.57 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.04 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.09 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 101.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 90.49 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.89 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 96.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 103.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 92.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 105.76 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 94.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 100.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 96.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.43 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 125.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 114.17 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 131.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 120.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 127.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 116.11 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.31 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 128.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 117.81 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 135.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 124.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 130.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 119.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.75 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.13 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.89 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.59 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.74 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.96 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.73 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.26 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.37 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.59 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 127.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.90 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 125.46 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 114.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 131.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 120.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 127.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 116.11 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 128.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 117.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 135.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 124.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 131.06 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 119.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 104.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 92.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.08 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.35 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.89 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.14 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.01 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.03 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.49 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.96 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.01 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.87 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.60 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 140.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.01 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.64 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.49 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.99 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.54 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.13 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.73 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 141.87 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.03 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 144.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 130.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.79 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.74 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 145.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.37 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.72 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.49 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 95.72 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 83.81 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.72 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 105.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 93.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 99.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 95.49 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 97.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 103.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 99.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.13 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.73 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.90 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.65 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.22 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.17 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.72 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.72 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.03 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.08 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.61 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.73 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.75 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 105.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 93.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 99.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 107.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 95.49 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 109.43 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 97.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 103.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 111.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 99.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.14 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 101.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 107.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 104.86 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 122.25 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 111.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 106.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.51 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.64 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.69 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.15 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.63 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.52 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.33 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.19 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.20 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.60 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.28 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.35 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.13 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.52 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 101.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 107.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 116.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 104.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 122.25 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 111.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 106.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 99.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.42 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.54 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.73 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.31 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.86 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.67 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.59 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.20 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.48 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.73 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.33 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.35 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.28 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.33 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.99 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.19 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.06 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.97 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.59 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 94.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.26 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.06 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.54 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.56 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.25 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 137.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 140.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.56 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.75 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 96.60 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 84.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 100.59 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 88.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.50 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 88.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 95.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 85.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 101.15 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 92.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 98.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 94.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.28 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.51 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 108.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.13 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.54 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.33 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.69 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.64 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.06 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.11 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.56 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.68 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.72 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.20 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.26 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 112.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 89.06 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 95.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 114.50 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 90.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 08-Jun-2026 17:43 113.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 08-Jun-2026 17:43 92.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 08-Jun-2026 17:43 118.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 08-Jun-2026 17:43 98.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 08-Jun-2026 17:43 115.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 08-Jun-2026 17:43 94.71 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 106.41 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 95.36 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.11 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.09 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.09 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.38 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.99 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.18 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.10 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.01 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.69 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.53 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.22 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.66 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.38 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.06 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.06 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.73 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.32 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.18 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.01 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.80 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.11 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.35 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.80 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.16 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.63 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.80 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.49 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.25 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.84 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.02 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.83 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.72 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.12 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.48 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.99 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.20 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.63 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.08 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.48 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.97 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 129.53 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 143.52 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 132.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.84 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.59 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.50 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 127.12 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.77 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.73 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.32 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.27 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.07 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.62 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.78 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 107.22 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.51 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.12 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.18 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.34 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 87.88 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.05 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.03 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.01 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.90 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.10 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 93.84 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.66 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.48 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.67 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.57 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.73 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.37 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 145.67 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 126.16 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.65 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.48 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.40 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.27 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.73 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.53 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.59 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.32 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.14 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.27 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.19 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 143.28 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 146.13 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 131.07 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.83 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.18 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.80 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.57 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.15 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.55 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.41 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.00 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.38 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.07 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 154.11 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 134.67 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 157.20 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 137.71 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.46 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.76 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.55 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.68 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.41 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.82 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.48 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.86 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.96 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.27 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 130.91 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.06 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.07 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 99.34 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.77 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.59 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 110.77 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.49 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 113.73 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.45 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 161.52 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 152.12 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 164.62 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 155.03 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.80 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 125.80 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 142.02 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.72 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.89 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.41 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.12 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.20 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 160.54 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 154.55 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 163.52 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 157.50 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 140.44 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 129.76 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 143.41 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 132.70 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.13 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.88 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.98 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 120.82 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.08 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.28 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.28 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 102.48 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 91.12 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 109.55 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 98.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.74 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 101.38 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 119.65 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 108.27 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.98 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.11 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.60 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.62 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.62 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.84 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.49 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.84 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.08 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.43 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.95 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.31 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 115.51 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.46 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.38 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.47 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.09 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.96 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 126.02 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.47 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.14 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.68 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 123.03 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.69 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 114.60 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.49 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.29 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.15 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 104.34 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.80 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDensePagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.48 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.52 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 117.21 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 136.44 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 124.22 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.04 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 114.94 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.32 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 121.43 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.50 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.53 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 131.78 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.91 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.29 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 115.66 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.36 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalPagedVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.16 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.37 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 89.47 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 111.89 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 96.48 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 105.77 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 90.18 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 112.97 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 97.27 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 116.03 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 100.48 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 122.47 KB
fmhaSm103aKernel_QkvFp16OFp16H80PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 107.51 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.05 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.31 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 133.46 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.20 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 121.38 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 105.93 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.66 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 112.95 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 118.14 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 102.32 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 124.41 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 109.34 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 127.44 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 111.77 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 134.98 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 118.91 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 120.93 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 106.12 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 128.47 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.27 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 117.81 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 103.16 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 125.36 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 110.18 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 138.59 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 122.12 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 144.90 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 128.63 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 132.80 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 116.74 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 139.21 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 123.25 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 129.55 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 113.26 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 08-Jun-2026 17:43 135.85 KB
fmhaSm103aKernel_QkvFp16OFp16H80PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 08-Jun-2026 17:43 119.76 KB
LICENSE 08-Jun-2026 17:43 11.09 KB
Artifactory/7.117.25 Server