Index of sw-kernelinferencelibrary-public-generic-local/134850621dbbd55ed6b0c3fa7c29b968136c05ef/fmha/trtllm-gen
Name Last modified Size
../
include/ 21-Apr-2026 16:04 -
checksums.txt 21-Apr-2026 16:05 1.70 MB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 99.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 96.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 76.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 100.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 69.92 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 61.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 67.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 59.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 69.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 69.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 98.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 68.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 59.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 65.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 57.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 68.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 67.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 95.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 88.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 65.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 58.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 63.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 55.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.92 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 75.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 63.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 72.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 61.35 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 77.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 108.17 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 99.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 73.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 61.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 70.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 75.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 97.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 70.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 67.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 57.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 112.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 98.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 103.97 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 90.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 99.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 85.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 112.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 106.03 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.42 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 104.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 96.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 85.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 78.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 83.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 76.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 84.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 100.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 81.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 79.17 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 97.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.97 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 88.40 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 76.72 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 109.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 87.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 84.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 90.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 81.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 88.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 79.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 86.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 77.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 72.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.40 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 69.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 81.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 67.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 60.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 65.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 58.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 68.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 69.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 67.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 88.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 79.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 65.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 58.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 63.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 56.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 68.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 66.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 67.77 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 65.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 85.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 77.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 63.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 56.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 61.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 54.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 73.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 99.31 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 90.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 73.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 61.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 70.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 72.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 70.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 97.43 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 71.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 68.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 57.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 69.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 70.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 68.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 95.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 86.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 68.31 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 57.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 65.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 55.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 154.63 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 133.90 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.67 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 151.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 140.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.86 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 150.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.33 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.04 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.75 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.71 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 118.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.34 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.23 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.23 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.65 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 121.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.43 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 113.74 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.12 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.95 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.08 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 127.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.71 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.12 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 126.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.85 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.01 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.72 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 124.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.47 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.75 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 178.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 174.46 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 176.18 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 173.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 166.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 160.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 159.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 171.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 167.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 169.50 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 166.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 159.18 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.06 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 152.16 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 168.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 166.43 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 163.03 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 156.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.96 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 182.42 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 178.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 170.00 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 147.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 165.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.47 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.20 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.16 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 169.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 162.44 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 141.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 157.37 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.53 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 168.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 159.35 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 154.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 275.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 270.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 244.72 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 258.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 236.14 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 253.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 259.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 254.77 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 228.59 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 243.35 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 221.79 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 238.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 251.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 247.25 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 223.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 236.22 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 216.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 231.54 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 279.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 275.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 247.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 229.76 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 240.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 224.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 263.24 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 258.83 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 231.85 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 213.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 224.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 209.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 255.86 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 251.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 226.45 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 208.57 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 219.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 203.68 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.92 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.37 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.76 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.11 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.95 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.57 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.53 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.68 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.49 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.21 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.79 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.96 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.00 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.80 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 101.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.21 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.54 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.22 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 95.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.50 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.40 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 111.15 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.27 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.44 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.61 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.01 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 122.83 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 118.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.94 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 101.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 110.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.90 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 154.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.53 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 134.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 145.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 141.54 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 135.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.78 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 141.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 132.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 126.68 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.67 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 146.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.61 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 145.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 132.94 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 147.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 143.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.00 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.85 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.74 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 227.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 222.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 196.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 210.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 187.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 205.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 210.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 206.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 180.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 194.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 190.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 203.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 175.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 187.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 168.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 182.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 231.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 226.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 199.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 181.64 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 192.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 176.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 215.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 210.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 183.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 165.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 160.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 207.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 202.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 171.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 155.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.54 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 113.21 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 101.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.85 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.74 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 97.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.67 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.93 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.68 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.11 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.43 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 98.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 111.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 98.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 104.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 128.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 153.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 134.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 141.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 135.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 126.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 122.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 123.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 102.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 121.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 136.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 130.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 131.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 115.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 135.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 135.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 121.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 80.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 104.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 93.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 134.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 126.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 97.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 144.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 151.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 134.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 129.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 140.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 115.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 134.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 112.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 151.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 143.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 140.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 142.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 140.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 132.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 149.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 128.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 152.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 148.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 137.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 153.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 142.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 127.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 146.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 132.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 142.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 150.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 95.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 135.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 128.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 83.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 78.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 130.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 102.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 102.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 89.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 98.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 98.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 108.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 105.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 95.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 119.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 103.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 150.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 141.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 155.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 153.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 169.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 148.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 142.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 128.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 110.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 153.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 138.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 129.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 104.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 140.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 155.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 128.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 132.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 112.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 126.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 118.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 100.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 134.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 120.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 115.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 98.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 133.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 147.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 107.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 94.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 131.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 133.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 129.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 133.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 111.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 157.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 160.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 160.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 128.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 168.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 152.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 155.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 145.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 137.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 158.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 174.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 154.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 138.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 149.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 159.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 146.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 140.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 130.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 137.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 136.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 145.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 160.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 136.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 124.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 108.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 132.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 115.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 137.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 119.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 152.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 113.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 135.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 118.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 116.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 100.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 155.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 121.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 128.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 105.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 89.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 103.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 113.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 157.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 158.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 165.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 153.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 160.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 149.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 168.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 157.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 156.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 157.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 141.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 130.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 116.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 131.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 117.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 163.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 157.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 166.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 160.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 132.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 146.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 135.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 119.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 133.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 165.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 163.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 162.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 162.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 164.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 155.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 165.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 154.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 162.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 152.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 162.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 145.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 141.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 143.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 131.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 117.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 128.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 135.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 232.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 246.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 225.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 144.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 155.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 209.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 197.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 232.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 222.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 198.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 190.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 167.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 194.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 209.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 185.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 209.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 171.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 161.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 136.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 126.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 193.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 171.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 134.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 171.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 193.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 177.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 157.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 144.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 149.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 237.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 250.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 229.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 149.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 159.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 214.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 201.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 236.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 226.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 202.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 198.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 213.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 191.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 134.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 189.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 213.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 164.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 141.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 162.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 183.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 176.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 175.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 160.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 197.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 181.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 163.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 154.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 108.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 93.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 99.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 93.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 118.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 117.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 115.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 110.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 131.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 95.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 79.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 114.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 74.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 117.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 117.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 118.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 108.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 103.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 111.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 108.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 106.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 128.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 77.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 94.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 134.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 84.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 121.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 106.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 106.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 89.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 93.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 76.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 92.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 88.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 72.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 134.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 101.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 108.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 110.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 130.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 116.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 121.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 119.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 112.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 119.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 121.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 142.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 134.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 128.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 100.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 141.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 134.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 119.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 95.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 81.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 99.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 77.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 118.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 112.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 94.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 138.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 130.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 124.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 97.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 80.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 108.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 115.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 92.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 101.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 112.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 98.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 97.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 94.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 80.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 89.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 76.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 103.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 164.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 167.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 181.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 151.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 161.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 163.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 131.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 150.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 151.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 165.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 140.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 159.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 151.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 151.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 167.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 178.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 165.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 143.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 159.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 147.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 148.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 121.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 145.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 144.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 147.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 133.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 161.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 122.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 135.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 114.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 125.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 108.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 136.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 139.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 113.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 131.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 102.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 103.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 175.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 184.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 159.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 165.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 175.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 162.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 180.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 161.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 131.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 147.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 119.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 162.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 170.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 182.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 150.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 162.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 159.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 151.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 146.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 156.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 166.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 141.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 134.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 165.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 129.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 143.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 145.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 131.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 111.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 147.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 144.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 156.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 150.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 151.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 144.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 156.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 134.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 136.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 124.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 108.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 150.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 150.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 190.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 158.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 200.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 167.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 183.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 181.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 160.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 145.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 195.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 187.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 155.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 148.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 170.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 155.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 139.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 138.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 126.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 148.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 143.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 125.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 114.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 162.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 173.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 124.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 163.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 148.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 176.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 160.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 143.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 157.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 143.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 96.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 128.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 107.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 141.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 144.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 148.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 155.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 104.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 138.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 121.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 128.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 149.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 145.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 154.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 168.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 158.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 168.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 210.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 170.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 224.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 184.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 204.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 165.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 145.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 200.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 197.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 219.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 214.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 189.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 175.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 165.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 168.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 148.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 160.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 193.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 191.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 166.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 151.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 137.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 123.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 151.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 137.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 169.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 165.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 152.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 162.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 132.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 137.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 141.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 178.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 155.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 193.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 169.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 124.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 178.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 165.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 158.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 195.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 181.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 174.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 161.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 148.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 143.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 131.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 171.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 158.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 151.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 151.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 135.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 129.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 133.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 163.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 178.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 163.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 119.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 152.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 180.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 168.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 140.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 123.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 156.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 143.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 146.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 143.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 127.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 150.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 156.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 157.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 164.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 165.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 151.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 172.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 162.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 144.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 150.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 152.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 200.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 201.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 249.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 192.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 271.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 214.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 242.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 185.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 162.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 147.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 237.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 236.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 195.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 179.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 264.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 259.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 220.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 202.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 200.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 195.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 227.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 228.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 186.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 172.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 137.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 121.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 168.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 151.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 131.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 189.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 195.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 146.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 130.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 159.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 162.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 206.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 174.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 228.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 196.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 198.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 167.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 206.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 193.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 161.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 230.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 216.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 201.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 183.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 166.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 154.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 196.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 185.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 169.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 153.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 129.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 163.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 157.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 131.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 141.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 188.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 166.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 210.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 188.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 180.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 159.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 128.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 138.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 131.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 188.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 170.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 153.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 212.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 198.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 193.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 147.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 167.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 161.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 146.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 144.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 125.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 138.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 150.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 123.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 118.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 180.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 184.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 163.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 188.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 186.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 174.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 170.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 158.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 94.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 109.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 96.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 94.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 107.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 136.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 111.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 128.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 111.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 115.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 73.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 84.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 124.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 94.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 110.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 95.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 112.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 72.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 83.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 107.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 71.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 139.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 108.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 125.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 120.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 119.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 112.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 120.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 141.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 116.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 136.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 133.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 81.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 135.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 126.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 112.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 98.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 98.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 92.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 95.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 81.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 108.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 109.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 104.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 91.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 103.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 126.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 147.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 123.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 114.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 99.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 135.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 121.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 106.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 104.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 96.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 101.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 87.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 103.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 105.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 90.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 79.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 109.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 146.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 106.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 128.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 106.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 144.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 95.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 119.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 113.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 98.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 124.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 131.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 95.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 107.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 102.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 88.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 104.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 91.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 103.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 105.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 92.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 98.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 99.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 124.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 116.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 103.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 113.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 146.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 133.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 119.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 111.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 128.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 144.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 146.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 134.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 186.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 179.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 199.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 183.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 164.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 149.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 171.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 164.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 146.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 183.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 167.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 151.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 158.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 177.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 177.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 190.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 184.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 203.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 154.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 134.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 155.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 125.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 168.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 188.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 169.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 169.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 162.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 182.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 163.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 151.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 102.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 106.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 93.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 130.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 109.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 88.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 85.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 104.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 87.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 119.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 105.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 101.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 105.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 110.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 116.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 122.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 114.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 96.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 121.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 108.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 135.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 116.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 115.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 88.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 110.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 115.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 94.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 80.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 95.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 100.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 141.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 101.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 118.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 126.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 132.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 103.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 105.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 108.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 97.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 152.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 154.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 142.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 142.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 108.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 131.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 143.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 103.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 142.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 104.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 146.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 160.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 120.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 130.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 148.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 109.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 114.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 109.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 129.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 130.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 139.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 104.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 112.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 141.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 117.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 159.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 177.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 159.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 122.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 107.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 109.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 105.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 108.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 122.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 162.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 154.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 113.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 109.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 121.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 116.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 102.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:05 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 103.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 146.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 151.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 134.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 108.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 95.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 128.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 122.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 111.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 97.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 89.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 111.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 135.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 124.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 21-Apr-2026 16:06 106.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 178.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 201.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 150.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 127.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 183.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 161.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 207.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 189.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 183.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 155.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 158.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 153.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 161.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 184.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 165.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 188.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 159.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 155.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 144.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 132.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 141.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 126.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 176.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 167.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 150.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 180.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 163.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 155.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 151.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 153.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 121.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 115.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 116.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 116.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 80.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 91.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 118.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 131.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 79.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 87.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 123.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 150.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 97.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 119.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 137.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 141.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 126.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 121.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 104.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 118.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 137.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 83.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 104.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 90.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 140.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 138.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 94.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 102.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 114.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 88.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 117.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 99.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 141.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 130.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 103.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 132.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 124.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 142.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 113.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 125.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 90.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 142.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 138.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 146.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 99.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 128.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 129.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 129.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 145.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 138.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 147.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 121.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 189.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 168.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 188.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 204.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 187.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 169.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 159.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 174.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 118.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 173.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 189.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 154.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 119.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 168.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 145.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 166.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 182.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 128.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 100.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 172.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 192.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 209.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 191.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 177.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 193.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 172.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 171.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 187.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 168.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 112.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 107.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 125.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 108.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 95.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 124.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 133.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 105.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 116.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 113.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 113.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 117.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 78.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 114.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 94.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 136.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 111.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 86.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 98.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 111.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 126.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 105.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 100.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 111.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 84.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 107.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 123.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 109.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 113.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 75.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 87.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 124.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 127.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 120.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 138.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 119.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 120.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 131.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 145.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:11 130.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 94.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 93.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 91.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 107.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 94.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:05 104.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 125.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 110.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 129.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 136.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 95.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 124.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 103.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 126.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 100.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 86.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 104.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:05 96.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 153.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 155.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 148.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 131.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 115.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 124.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 115.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 137.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 95.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 105.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 148.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:05 141.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 150.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 133.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:05 130.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 148.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 131.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 151.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 134.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 149.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 151.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 193.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 161.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 165.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 154.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 178.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 191.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 180.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 200.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 181.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 131.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 146.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 165.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 185.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 165.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 139.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 159.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 158.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 136.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 105.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 88.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:05 103.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:05 92.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 86.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 87.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 91.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 103.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 105.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 102.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 88.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 105.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 107.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 97.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 129.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 128.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 146.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 146.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 180.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 200.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 184.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 164.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 164.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 184.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 167.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 149.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 165.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 158.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 120.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 177.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 191.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 170.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 184.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 166.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 204.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 162.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 176.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 154.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 169.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 150.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 189.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 162.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 182.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 163.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 151.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 103.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 125.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 91.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 88.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 104.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 87.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 90.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 100.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 82.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 81.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 97.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 113.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 95.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 154.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.46 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 143.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.22 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 150.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.79 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.03 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 80.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.99 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.79 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.50 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 118.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.07 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 154.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 141.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 133.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.14 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 140.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.96 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 135.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.22 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.14 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.37 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 152.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 137.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 160.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 140.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 148.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 133.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 149.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 153.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 150.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 138.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 106.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 155.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 133.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 96.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 152.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.60 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 139.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 124.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 135.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 141.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 129.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 101.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 90.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 93.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 151.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 169.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 142.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 156.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 169.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 149.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 143.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 117.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 141.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 148.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 132.93 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.43 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 140.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 156.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 112.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 119.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.82 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.18 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.43 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 148.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 109.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 92.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 121.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.00 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 106.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 96.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 99.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 132.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 149.00 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 135.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 159.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 174.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 147.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 153.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 161.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 157.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 168.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 148.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 157.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 147.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 138.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 158.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 174.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 149.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 160.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 147.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 141.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.50 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 138.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 137.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 114.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 98.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 158.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.40 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 146.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 121.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 152.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 137.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.72 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 142.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 131.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.08 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 145.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 161.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 142.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 122.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 136.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 116.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 134.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 125.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 116.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.40 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 132.04 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.94 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 136.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.14 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 153.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 136.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 127.67 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 104.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 101.25 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.39 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.65 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.82 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 159.90 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 159.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 158.01 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 166.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 157.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 155.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 161.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 151.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 169.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 159.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 157.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 158.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.37 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.99 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.83 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.22 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 143.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 138.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.18 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 142.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 131.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 139.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.76 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.82 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.04 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.69 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 125.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.56 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 124.51 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.42 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 164.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 158.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 167.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 161.69 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 133.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 136.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 166.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 165.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 164.24 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 163.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 165.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 157.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 166.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 156.07 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 163.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 153.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 164.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 149.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 140.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 143.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 146.33 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 142.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 147.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 144.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.48 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 128.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 122.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 232.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 247.08 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 225.26 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 155.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 210.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 233.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 222.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 198.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 190.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 147.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 168.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.10 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 137.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 210.44 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 186.87 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 130.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 141.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 185.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 173.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 210.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 171.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 161.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 136.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 133.01 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 126.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 179.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 171.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 172.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 156.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 194.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 177.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 157.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 144.51 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 150.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 126.50 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 237.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 251.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 229.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 149.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 143.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 214.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 202.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 237.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 226.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 202.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 194.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 152.87 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 135.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 174.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 151.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 142.10 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.11 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 214.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 191.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 134.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 146.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 190.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 177.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 214.93 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 198.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 178.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 164.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 141.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 138.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 130.85 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 183.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 199.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 175.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 128.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 139.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 176.00 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 160.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 198.83 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 181.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 163.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 147.96 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 134.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 155.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 131.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 123.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 109.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 94.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 115.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.60 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 113.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 124.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 139.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 120.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 115.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 128.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 111.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.51 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 130.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 120.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 97.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 90.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 119.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 105.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 97.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 107.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 84.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 104.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 111.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 124.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 91.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 105.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.01 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 127.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 94.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.11 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.18 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.11 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 115.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 106.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 115.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 109.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 83.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 117.39 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 103.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 122.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 107.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 106.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 89.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 102.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 85.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 110.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 93.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 88.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 130.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 138.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 121.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 129.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 105.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 126.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 118.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 115.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 123.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 114.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 101.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 121.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 108.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 125.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 108.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 136.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 129.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 145.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 103.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 126.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 90.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 127.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 117.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 127.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 135.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 119.08 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 129.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 114.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 119.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 104.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 133.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 118.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 99.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 113.89 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 136.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 121.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 100.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 83.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 110.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 92.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 120.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 102.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 78.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 134.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 125.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 121.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 110.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 142.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 102.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 124.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 89.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 125.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 113.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 123.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 113.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 109.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 133.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 100.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 116.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 88.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:04 126.64 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 111.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 114.90 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 100.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 129.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 114.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:04 114.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 96.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 95.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 141.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:04 133.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 127.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 119.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 81.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 109.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 118.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 99.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 93.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 132.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 122.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 119.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 108.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 141.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 92.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 101.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 122.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 88.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 123.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 111.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 120.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 112.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 107.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 131.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 91.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 99.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:04 115.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:04 125.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 110.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 112.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 98.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 127.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 112.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 112.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 94.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 107.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 132.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 126.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 117.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 97.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 80.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 108.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 90.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 21-Apr-2026 16:06 116.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 21-Apr-2026 16:06 98.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 92.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 76.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 100.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 97.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 70.33 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 62.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 68.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 60.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 74.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 69.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 100.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 90.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 68.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 60.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 66.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 58.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 68.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 71.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 68.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 97.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 88.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 66.33 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 58.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 64.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 56.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 75.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 78.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 76.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 64.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 73.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 61.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 73.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.01 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 96.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 73.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 61.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 70.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 59.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 76.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 74.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 104.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 71.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 59.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 68.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 57.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.53 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 101.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 85.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 101.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 98.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 99.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 97.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 95.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 85.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 92.83 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 83.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 91.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 93.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 90.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 98.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 86.58 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 79.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 84.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 76.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 90.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 86.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 85.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.67 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 82.58 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 74.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 79.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 72.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 104.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 101.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 102.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 99.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 98.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 87.25 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 95.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 85.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 98.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 94.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 96.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 93.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 92.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 79.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 89.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 77.50 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 93.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 89.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 91.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 88.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 88.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 75.38 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 85.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 73.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 91.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 83.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 89.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 80.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 87.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 78.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 72.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 70.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 70.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 91.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 83.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 68.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 60.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 66.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 58.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 70.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 68.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 69.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 68.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 89.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 80.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 66.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 58.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 64.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 56.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 69.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 67.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 68.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 66.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 87.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 78.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 64.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 57.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 62.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 55.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 75.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 73.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 74.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 72.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 98.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 74.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 61.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 71.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 73.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 71.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 72.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 70.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 96.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 87.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 71.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 59.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 68.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 57.50 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 72.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 69.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 71.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 69.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 93.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 85.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:06 68.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:06 57.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 21-Apr-2026 16:04 65.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 21-Apr-2026 16:04 55.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.63 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 113.17 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.19 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 126.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.14 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.31 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.77 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.06 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 123.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.71 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 134.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 151.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 138.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 132.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.89 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 116.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.78 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.31 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 106.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 108.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 95.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.43 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 129.17 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.47 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 144.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 133.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.39 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.49 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 155.77 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 137.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 159.22 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 140.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.37 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 115.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.75 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.58 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.11 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 105.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.11 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 166.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 155.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 169.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 157.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 144.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 128.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 147.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.48 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 130.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 164.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 157.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 167.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 160.51 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 144.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 132.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.08 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 134.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.71 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.61 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 98.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.32 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.68 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.79 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 117.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.72 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.07 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 138.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 104.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.61 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.10 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.97 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 112.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 96.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 100.78 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.67 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.21 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 138.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.03 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 134.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.97 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 96.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 101.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 88.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 104.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 91.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.95 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.54 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 111.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.84 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.84 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.54 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.00 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.84 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.00 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.36 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.50 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.84 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.28 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.47 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 102.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 88.99 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 104.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 91.01 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 103.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 92.11 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 110.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.28 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.36 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.50 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.21 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.42 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.49 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.11 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.63 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 144.06 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.03 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 130.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.72 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 103.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.08 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 103.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.93 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.50 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.14 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 103.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.30 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.67 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.69 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.81 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 110.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 114.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 111.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 98.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 124.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.03 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.74 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.32 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 140.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.61 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.07 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 138.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 102.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.83 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 91.58 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 104.03 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 106.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.99 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 119.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.00 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.67 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 121.15 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.73 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.65 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 141.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 135.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.69 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.04 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 111.69 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.90 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.32 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 98.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.71 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 130.21 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.19 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.94 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.92 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.31 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.13 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 114.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.13 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.76 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.82 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 135.34 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.34 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.46 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.83 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 111.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.42 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.28 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.63 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 129.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.41 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.36 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.82 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.00 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.01 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.34 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.17 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.90 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.94 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.50 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.97 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 102.57 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.60 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.89 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.65 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.43 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.10 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 137.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.83 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 145.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 134.25 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.17 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.54 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 106.93 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 110.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.26 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 111.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 96.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 116.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.22 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.32 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.10 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.25 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 135.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 128.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.03 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.35 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.43 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 130.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.83 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 124.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 121.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.03 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.54 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 105.26 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 132.37 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 151.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 134.39 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.33 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.08 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 131.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.14 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 118.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 149.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 132.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 152.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 135.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.50 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 119.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 103.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 107.08 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 103.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.35 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.22 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.75 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.26 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.65 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 106.65 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.14 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.57 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.64 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 118.69 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.59 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.31 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 107.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.01 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 133.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 132.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.48 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 96.22 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.08 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 110.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.75 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 110.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 127.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.63 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 122.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.42 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.72 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.36 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.82 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 122.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.87 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.47 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 101.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 89.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 91.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 103.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 89.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 105.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 92.36 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.21 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 99.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.63 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:06 112.08 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.22 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.64 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.68 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.70 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 120.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.42 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 123.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.08 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:06 117.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.48 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.14 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.75 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.26 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.15 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.21 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.72 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 111.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 97.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 99.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 112.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 98.06 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.59 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.60 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.67 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.87 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.89 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 144.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.13 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 131.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 127.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.31 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 104.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 104.75 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 91.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 108.64 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.63 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.14 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.26 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.40 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 119.57 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.57 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.47 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.03 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 121.93 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 107.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.61 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.52 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.33 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.99 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 113.76 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 100.77 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.50 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 114.30 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.79 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.33 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.71 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.33 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.41 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.16 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.40 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.27 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.45 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.12 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.35 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.53 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 135.91 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.20 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.58 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.65 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.13 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.35 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 137.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.98 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 141.65 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 128.96 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.23 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.31 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 123.17 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.59 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.84 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 136.00 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 152.39 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 139.41 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.80 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 147.34 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 133.49 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 140.03 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 126.20 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 143.57 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.73 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 107.26 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 90.97 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 110.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 94.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 106.96 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 93.98 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 110.06 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 96.93 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 117.68 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 101.56 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.68 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 104.84 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.19 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.39 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 149.40 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 130.30 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.24 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 111.50 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.60 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 114.41 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.05 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.55 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.18 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.59 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 146.11 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 132.15 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.98 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 135.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.14 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.33 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.43 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.15 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.52 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.30 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 125.62 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 156.90 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 138.70 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 160.36 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 141.74 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 142.05 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.79 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.44 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.83 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.50 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.97 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.57 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.89 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.13 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.24 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.96 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.24 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.52 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 105.27 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 118.80 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.34 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 167.37 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 156.14 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 170.59 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 158.94 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.13 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 129.84 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.36 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 132.62 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 131.61 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 118.44 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.97 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.23 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 165.31 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 158.82 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 168.65 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 161.64 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 145.37 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 133.29 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 148.22 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 136.12 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 132.19 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 121.91 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 135.60 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 124.73 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.35 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 115.70 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 102.01 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 109.66 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 95.96 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 116.30 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 103.05 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 120.02 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 106.09 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.27 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 113.07 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.92 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.02 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.47 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.41 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.74 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.48 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.62 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 122.90 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 108.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 128.94 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 115.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 126.95 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 112.87 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 133.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.38 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 124.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 110.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 130.91 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 117.58 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 123.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 109.05 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 129.34 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 116.18 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 138.36 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 122.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 144.31 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 130.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 136.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 120.84 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 141.33 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 127.22 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 134.46 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 119.30 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 21-Apr-2026 16:04 139.78 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 21-Apr-2026 16:04 125.80 KB
LICENSE 21-Apr-2026 16:06 11.09 KB
Artifactory/7.117.25 Server