Index of sw-kernelinferencelibrary-public-generic-local/84ee55b93ba92b1f64bbf727fd1cf38d9d3058a7/fmha/trtllm-gen
Name Last modified Size
../
include/ 16-Apr-2026 13:50 -
checksums.txt 16-Apr-2026 13:51 1.82 MB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 99.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 96.45 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 100.79 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 69.92 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 61.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 67.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 69.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 69.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 98.54 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 68.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 65.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 57.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 68.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 67.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 95.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 65.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 58.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 63.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 55.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.92 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 75.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 63.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 72.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 61.35 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.17 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 73.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 61.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 70.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 70.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 70.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 67.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 57.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 103.97 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 90.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 85.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.03 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.42 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 85.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 83.30 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 100.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 81.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.99 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 79.17 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.38 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.97 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.12 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.40 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.72 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.57 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 87.78 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 84.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.22 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.39 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.19 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.14 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.06 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.02 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.33 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.05 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.33 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.16 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.84 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.89 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.85 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.41 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.04 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.21 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.50 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.66 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.46 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.13 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.82 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.29 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.91 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.34 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.09 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.26 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.94 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 90.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 81.96 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 88.68 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 79.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 86.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 77.59 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 70.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.40 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 69.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 90.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 81.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 67.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 60.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 65.67 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 58.24 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 70.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 68.32 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 69.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 67.49 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 88.00 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 79.60 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 65.95 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 58.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 63.80 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 56.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 68.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 66.69 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 67.77 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 65.87 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 85.75 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 77.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 63.70 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 56.52 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 61.55 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 54.62 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.27 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 99.31 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 90.86 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 73.56 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 61.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 70.53 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.07 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 70.37 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 97.43 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.61 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 71.18 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 59.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 68.15 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 57.10 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.47 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 69.44 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 70.65 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 68.74 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 95.01 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 86.48 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 68.31 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 57.25 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 65.28 KB
fmhaSm100aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 55.10 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.24 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.28 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.61 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.28 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.77 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 104.38 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 125.55 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 110.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 106.29 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.29 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.19 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.30 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.31 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.74 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.21 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.06 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.54 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.92 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.54 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.00 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.52 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.90 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.57 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.07 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.02 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.56 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.53 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.41 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 131.37 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.05 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.26 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.82 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 139.97 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.58 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.28 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.43 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.05 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.86 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 104.51 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 125.68 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 110.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.28 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 106.29 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.40 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.34 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.42 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.74 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.76 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.29 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.74 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.02 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.08 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.70 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.82 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.92 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.65 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.75 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.85 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.22 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.32 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.02 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.59 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.81 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.38 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.86 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.78 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.31 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.34 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.91 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.94 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.00 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.76 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.32 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.14 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.86 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.58 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.14 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.43 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.35 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.56 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 148.57 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 131.63 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.26 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.87 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.08 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.38 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.15 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.79 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 148.94 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 132.65 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.21 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.57 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.27 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.88 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.90 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.01 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.93 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.95 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 102.05 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 127.26 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 105.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 123.79 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 103.77 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.33 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.89 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.82 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.76 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.47 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.88 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.86 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.82 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.15 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.41 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.36 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.73 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.34 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.87 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.96 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.19 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.95 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.09 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.31 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.23 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.19 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.77 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.52 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.14 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.39 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.23 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.98 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.59 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.95 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.67 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.57 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.25 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.13 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.21 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.77 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 139.16 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.45 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.67 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.24 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.62 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.80 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.07 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 102.18 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 127.38 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 105.75 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 123.91 KB
fmhaSm100aKernel_QkInt8VE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 103.89 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.63 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.90 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.67 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.86 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.33 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.04 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.75 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.71 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.34 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.23 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.23 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.65 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.43 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.74 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.12 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.95 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.08 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.71 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.12 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.85 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.01 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.72 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.47 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.75 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.46 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.18 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 166.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.50 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.18 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.06 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.16 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.43 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.03 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.96 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 182.42 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 170.00 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 165.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 175.47 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.20 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.16 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 162.44 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.37 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.53 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.35 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 154.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 279.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 273.07 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 275.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 270.91 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 244.72 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 258.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 236.14 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 253.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 262.82 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 256.80 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 259.05 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 254.77 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 228.59 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 243.35 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 221.79 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 238.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 255.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 249.42 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 251.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 247.25 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 223.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 236.22 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 216.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 231.54 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 283.40 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 277.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 279.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 275.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 247.99 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 229.76 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 240.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 224.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 267.13 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 260.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 263.24 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 258.83 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 231.85 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 213.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 224.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 209.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 259.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 253.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 255.86 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 251.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 226.45 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 208.57 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 219.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 203.68 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.92 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.31 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.37 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.76 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.11 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.95 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.57 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.53 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.62 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.51 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.19 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.02 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.68 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.48 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.94 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.49 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.70 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.38 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.52 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.21 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.79 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.96 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.29 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.84 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.00 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.80 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.26 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.66 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.21 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.54 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.22 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.10 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.98 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.50 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.28 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.97 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.40 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.41 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.15 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.27 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.55 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.56 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.44 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.87 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.32 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.88 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.61 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.73 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.01 KB
fmhaSm100fKernel_QBfloat16KvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.83 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.81 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.58 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.94 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.70 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.22 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.90 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.53 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.06 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.54 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.28 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.78 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.68 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.65 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.67 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.61 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.52 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.94 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.00 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.87 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.85 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.74 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 230.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 224.32 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 227.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 222.16 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 196.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 210.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 187.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 205.33 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 214.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 208.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 210.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 206.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 180.48 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 173.30 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 190.18 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 207.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 200.80 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 203.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.76 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 175.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 187.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 168.13 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 182.91 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 235.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 228.51 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 231.38 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 226.34 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 199.99 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.64 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 192.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 218.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 212.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 215.12 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 210.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 183.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 176.47 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.72 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 211.50 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 204.86 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 207.73 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 202.95 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.57 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.19 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.44 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.40 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.59 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.75 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.03 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.54 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.05 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.21 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.20 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.88 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.77 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.85 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.26 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.31 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.39 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.07 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.08 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.37 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.04 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.41 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.62 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.66 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.74 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.55 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.84 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.63 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.23 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.79 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.24 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.67 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.93 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.68 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.09 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.11 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.42 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.43 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.98 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.10 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.97 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.25 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.27 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.02 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.56 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.71 KB
fmhaSm100fKernel_QFp16KvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.86 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.41 KB
fmhaSm100fKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 153.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 132.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 152.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 148.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 137.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 153.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 146.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 132.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 102.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 89.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 155.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 169.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 148.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 142.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 155.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 131.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 133.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 157.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 160.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 155.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 137.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 158.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 174.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 159.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 146.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 140.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 160.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 155.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 89.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 157.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 165.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 156.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 153.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 160.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 168.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 157.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 156.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 157.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 163.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 157.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 166.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 160.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 132.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 146.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 135.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 165.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 163.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 162.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 162.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 164.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 155.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 165.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 162.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 152.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 162.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 237.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 252.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 228.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 232.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 246.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 225.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 209.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 197.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 232.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 222.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 198.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 190.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 167.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 216.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 189.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 209.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 185.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 209.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 200.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 193.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 193.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 241.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 257.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 232.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 237.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 250.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 229.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 214.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 201.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 236.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 226.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 202.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 173.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 203.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 220.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 213.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 189.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 213.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 162.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 188.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 205.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 175.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 197.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 163.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 154.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 139.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 94.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 80.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 164.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 163.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 151.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 167.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 148.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 161.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 175.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 184.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 159.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 175.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 162.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 180.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 161.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 162.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 170.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 182.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 150.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 162.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 159.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 146.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 156.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 165.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 144.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.63 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.66 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 190.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 181.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 160.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 148.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 176.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 155.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 148.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 163.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.17 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.64 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.20 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.13 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.79 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.94 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.60 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 212.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 206.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 154.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 168.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 158.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 168.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 210.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.49 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 204.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.33 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 200.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 197.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 173.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 165.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 168.85 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 148.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 160.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 193.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 166.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.10 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 169.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 165.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 152.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 162.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 180.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.96 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 137.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.77 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 141.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 158.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 148.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 131.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 151.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 151.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 135.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 129.78 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 133.68 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.90 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.59 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.69 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 163.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.62 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 140.08 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 123.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.57 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.18 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.58 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 143.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 127.91 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.14 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.71 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 150.29 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 156.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 172.06 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 162.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 144.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 253.22 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 195.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 244.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 187.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 200.98 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 201.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 249.48 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 192.15 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 242.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 185.26 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 237.44 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 236.73 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 195.92 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 200.37 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 195.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 227.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 228.46 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 186.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.83 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 189.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 195.43 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.11 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.76 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 209.82 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 201.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.24 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.93 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.04 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 159.61 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 162.99 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 206.34 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.89 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.97 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 206.03 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 193.47 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 166.19 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.00 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 196.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 185.45 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 169.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.74 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.25 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 163.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 157.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.54 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.41 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.88 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.12 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.02 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.70 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.42 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 188.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.75 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 180.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.86 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.72 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.51 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 188.28 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 175.53 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 170.95 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.67 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 147.01 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.81 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.52 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.39 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 161.84 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.23 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.50 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.36 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 150.35 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.55 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.38 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.07 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.65 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.87 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.21 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.32 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 180.31 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 175.56 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.27 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 176.05 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.30 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 186.40 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 174.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 170.80 KB
fmhaSm100fKernel_QkvBfloat16OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 114.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 106.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 102.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 95.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 81.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 115.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 106.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 102.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 104.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 101.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 87.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 144.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 102.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 146.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 192.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 179.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 199.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 164.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 164.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 183.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 158.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 177.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 197.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 190.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 184.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 203.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 172.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 155.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 182.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 168.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 188.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 162.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 182.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 151.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 100.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 93.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 99.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 95.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 88.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 104.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 87.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 94.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 80.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 118.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 93.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 99.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 95.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.42 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 152.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 142.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.81 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.21 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk320HV256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.89 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.85 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.63 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 104.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.92 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.00 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.18 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.33 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.67 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta128PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.74 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 112.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 122.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 107.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 124.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 109.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.44 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.51 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.65 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 105.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 108.36 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 113.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.14 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.20 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.13 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 116.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 102.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.97 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.83 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.75 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.17 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 100.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 103.50 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.79 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.22 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.72 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 108.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 95.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.41 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.66 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 111.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.88 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 97.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.87 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 111.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.35 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Persistent2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 124.03 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128Static2CtaKeepsAbForGen.cubin 16-Apr-2026 13:51 106.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512HVPerCta256PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 182.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.04 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.19 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.90 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.70 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 183.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.37 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 161.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 155.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 173.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.69 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.34 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.61 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.11 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 153.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.82 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.73 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP16VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.07 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.46 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.49 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.71 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.06 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.95 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.28 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.39 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 165.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.29 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.23 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 155.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.45 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 144.94 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.08 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.58 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.10 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP32VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.93 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.80 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.01 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.68 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.31 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.54 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.24 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.76 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.32 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.02 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.38 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.96 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.05 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.43 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.27 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.60 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.99 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.25 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.53 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.30 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.40 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.26 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ16Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.47 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.64 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.57 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.52 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.56 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvDenseP64VarSeqSkipsSoftmaxQ8Kv64StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.55 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.78 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.48 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.91 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.86 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.16 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.98 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 153.59 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.84 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.15 KB
fmhaSm100fKernel_QkvE4m3OBfloat16HQk576HV512PagedKvStaticTokenSparseP1VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 101.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 123.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 108.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 119.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 103.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 101.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 123.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 108.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 119.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 103.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 103.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 88.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.06 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 85.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 146.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 145.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 147.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.82 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.79 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 195.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 175.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 189.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 188.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 204.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 187.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 169.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.63 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 173.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 189.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 154.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.60 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 166.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 182.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.37 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 200.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 192.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 209.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 185.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.53 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 177.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 193.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 175.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 163.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.23 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 187.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 108.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 94.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 114.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 100.88 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 110.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 96.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.75 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.40 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.65 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.33 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.28 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.14 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.39 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.94 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.21 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.12 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.96 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.85 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.97 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.32 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.83 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.19 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.71 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.03 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.58 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.18 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.95 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.70 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.16 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.87 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.72 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.43 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.90 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.46 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.31 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.61 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.56 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.74 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.84 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.64 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.11 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.52 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.17 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.54 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.76 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.05 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.69 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.09 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.01 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.81 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.02 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.22 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.42 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.44 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.93 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.29 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.78 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.49 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.86 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.80 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.13 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.77 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.27 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.45 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.55 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.30 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.36 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.35 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.73 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.24 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.08 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.91 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.10 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.57 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.20 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.62 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.98 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.07 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.99 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.48 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.59 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.50 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.15 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.67 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.04 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.26 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.38 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.68 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.89 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.47 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.66 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 108.25 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 94.92 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 114.41 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 101.00 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 110.34 KB
fmhaSm100fKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 96.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 114.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 134.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 120.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 130.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 115.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 87.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 131.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 139.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 128.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 114.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 134.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 120.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 130.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 116.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.28 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 103.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 100.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 86.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 104.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 153.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 155.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 146.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 148.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 150.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 133.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 150.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 134.37 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 149.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 151.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 193.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 176.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 193.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 163.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.06 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 161.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.17 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 154.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 172.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 141.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 145.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 180.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 200.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 172.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 131.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 146.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 165.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 185.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 139.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.94 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.83 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 159.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.46 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 100.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 111.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 98.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 104.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 113.90 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 99.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.97 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.54 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.92 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.35 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.56 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.96 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.22 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.95 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.57 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 88.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.39 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.86 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.77 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.13 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.70 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 103.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 86.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.32 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 89.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.51 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.66 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.42 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.30 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.61 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.14 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.50 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.40 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.72 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.26 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.47 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.08 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.76 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.12 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.60 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.34 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.98 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.59 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.04 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.24 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.07 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.03 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.69 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.52 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.81 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.48 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.21 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.91 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.01 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.19 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.65 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.82 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.09 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.15 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.55 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.89 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.44 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.33 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.53 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.62 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.31 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.58 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.23 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.63 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.75 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.99 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.64 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.05 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.43 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.88 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.84 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.87 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.78 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.49 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.73 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.16 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.29 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.68 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.74 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.45 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.38 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.80 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.67 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.79 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.00 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.41 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.85 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 111.93 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 98.10 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 118.11 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 104.18 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 114.02 KB
fmhaSm100fKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 100.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 111.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 114.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 106.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 116.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 102.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 97.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 73.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 87.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 71.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 138.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 140.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 137.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 115.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 100.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 106.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 117.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 102.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 87.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 104.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 102.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 91.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 102.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 88.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 105.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 107.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 95.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 98.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.08 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 128.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 123.68 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 126.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 126.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 115.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 146.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 193.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.15 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 180.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 200.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 184.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 164.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 164.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 184.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 167.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 158.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 178.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 120.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 181.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 172.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 191.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 170.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 184.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 204.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 172.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 134.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 166.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 162.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 154.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 169.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 189.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.70 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 159.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 176.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 150.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 129.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 162.81 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 182.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 163.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 151.46 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 115.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 102.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 101.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 119.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 93.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 99.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 120.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 95.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 88.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 105.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 92.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 116.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 103.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 125.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 91.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.33 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 105.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 94.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 91.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.82 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.57 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 106.07 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 88.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.80 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.39 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 89.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.90 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.60 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.65 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.89 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 100.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.22 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 104.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 87.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 103.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.94 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 108.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 90.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 88.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 72.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.61 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 100.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.55 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 102.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.21 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.06 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 102.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 106.50 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.52 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.14 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.18 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.47 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.23 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.88 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.35 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.10 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.77 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.42 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.69 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.76 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 100.36 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 82.54 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.11 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 78.13 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.83 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.09 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.78 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.37 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.40 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.91 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.20 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.32 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.41 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.38 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.24 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.19 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 111.45 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 93.92 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.17 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.97 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.28 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 98.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 81.26 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.53 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.93 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 115.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 97.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 93.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 76.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.30 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.84 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.67 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.74 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.27 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.34 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.25 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.62 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.87 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.72 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.86 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.66 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.71 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.16 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.79 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.31 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.96 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.99 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.85 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 110.05 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.01 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 124.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.02 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 110.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.51 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.64 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.00 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.04 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.43 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.75 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 97.58 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.48 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.44 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.29 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 113.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 95.73 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.63 KB
fmhaSm100fKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.95 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:51 119.12 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:51 93.49 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.59 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:51 99.56 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:51 121.03 KB
fmhaSm100fKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:51 95.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 118.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 110.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 129.43 KB
fmhaSm100fKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.46 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 124.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.49 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.26 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 130.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 143.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 102.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 84.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 95.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.58 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.22 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 150.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 123.79 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.32 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.03 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 113.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 80.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 121.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 92.99 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 75.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.79 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.50 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 128.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 118.85 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.07 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 118.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 90.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 74.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 154.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.44 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 133.17 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.14 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 140.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.96 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 135.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.22 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.14 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 141.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.65 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 135.28 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 121.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 127.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 113.37 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 132.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 152.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 137.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 160.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.81 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.77 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.29 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.74 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.11 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.52 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.54 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 148.31 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 133.78 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.63 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.24 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.33 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.55 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.95 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.10 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.69 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 153.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 150.18 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 138.72 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.75 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 106.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 88.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 115.64 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.40 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.53 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 155.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.25 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.05 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 139.57 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.35 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.56 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.87 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 91.01 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.86 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.62 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.82 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.30 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.71 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.76 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.84 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 133.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 103.80 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 85.09 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 95.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 126.94 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 110.90 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 96.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 79.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.83 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.23 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 152.48 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.73 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.12 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 90.41 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.68 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.60 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.13 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.70 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.91 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.36 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 139.45 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.20 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 124.88 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.93 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.66 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.61 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 135.16 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.51 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 120.27 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.02 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.34 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 141.59 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 129.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 126.04 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 101.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.67 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.98 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.19 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.15 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.08 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 94.38 KB
fmhaSm100fKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 77.97 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 112.89 KB
fmhaSm100fKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 99.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 107.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 90.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 110.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 94.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 106.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 93.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 109.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 120.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 169.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 142.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 156.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 164.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.97 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 127.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.70 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 169.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 142.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 149.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 143.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 130.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 117.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.25 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 109.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 94.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.69 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 141.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 132.93 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.43 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.87 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 111.51 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 105.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 140.73 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 122.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 156.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.44 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 130.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 114.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 112.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 92.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 119.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.82 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 86.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 129.29 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.61 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.18 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 135.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 113.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 97.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.54 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 109.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 125.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.43 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 121.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 116.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 105.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 99.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 133.58 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 148.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 109.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 92.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 121.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 106.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.00 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 108.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 106.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 107.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 100.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 112.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 96.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 99.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 83.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.99 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 132.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 149.00 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 135.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 122.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 109.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 125.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 112.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 159.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 147.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 153.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 161.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 128.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.78 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 157.74 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 168.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.79 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 148.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 115.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.37 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.53 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 157.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 147.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 138.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.07 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 158.31 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.88 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 174.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.86 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.77 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 149.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.95 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 160.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 147.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 141.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 122.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.50 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 138.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 137.49 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 114.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 98.09 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.89 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 158.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.40 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.26 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 146.57 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.12 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 121.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 134.38 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.33 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.71 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.84 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 121.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 137.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.39 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.52 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 117.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.72 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 102.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 142.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 131.90 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 122.80 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 113.08 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 145.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.16 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 161.22 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 142.36 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 122.28 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 104.17 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.20 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 116.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 145.46 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 134.96 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 125.85 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 116.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 116.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 96.40 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 132.04 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 108.75 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.47 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 108.30 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 89.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.94 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.32 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.62 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.41 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 107.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.27 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.34 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 101.45 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 136.24 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.01 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.03 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.35 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 104.76 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 114.14 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.55 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 123.56 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 116.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 107.13 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.64 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 119.81 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 153.23 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.02 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 114.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 98.11 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 128.65 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 136.59 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.19 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 119.91 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.15 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 111.60 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 93.63 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 127.67 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 105.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 101.42 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 104.05 KB
fmhaSm100fKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 87.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 115.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 101.25 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.39 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 114.65 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 104.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 117.82 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 107.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 159.90 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 159.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 158.01 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 166.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 157.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 155.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 161.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 151.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 169.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 159.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 157.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 158.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 145.37 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.99 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.83 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 133.22 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 143.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 128.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 138.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.18 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 142.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 131.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 131.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 139.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.76 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.82 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.04 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 120.69 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 130.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 125.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 117.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.56 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 124.51 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 108.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 133.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.42 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 118.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 127.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 111.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 164.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 158.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 167.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 161.69 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 144.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 133.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 147.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 136.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 131.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 120.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:51 134.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:51 123.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 166.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 165.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 164.24 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 163.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 165.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 157.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 166.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 156.07 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 163.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 153.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 164.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 149.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 140.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 143.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 138.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 146.33 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 130.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 142.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 127.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 147.14 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 136.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 144.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.48 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 128.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 123.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 132.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64MultiCtasKvGmemSepVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 125.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 119.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 129.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 114.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 136.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 122.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:51 132.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512HVPerCta256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:51 117.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 237.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 253.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 228.18 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 147.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.24 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 140.56 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 232.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 247.08 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 225.26 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 155.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 139.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 210.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 233.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 222.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 198.61 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 190.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 147.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 131.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 168.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.10 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 137.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 123.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.86 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 216.76 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 189.73 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 148.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.28 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 210.44 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 186.87 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 130.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.35 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 185.79 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 173.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 210.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 171.98 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 161.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 136.62 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.06 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.55 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 133.01 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 126.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 110.29 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 183.77 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 201.88 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 174.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 141.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 120.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 179.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.68 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 171.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 124.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 134.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 118.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 172.03 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 156.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 194.66 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 177.49 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 157.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.51 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 129.52 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 111.43 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 150.32 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 126.50 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 119.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 103.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 241.84 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 257.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 232.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.70 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 165.95 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 144.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 237.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 251.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 229.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 149.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 160.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 143.12 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 214.40 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 202.34 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 237.71 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 226.46 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 202.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 194.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 152.87 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 135.41 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:51 174.58 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 151.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 142.10 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 127.11 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 203.26 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 221.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 193.89 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 138.53 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 152.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 131.10 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 198.80 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 214.57 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 191.02 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 134.78 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 146.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 128.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 190.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 177.20 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 214.93 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 198.21 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 178.38 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 164.72 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 141.74 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 122.27 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 163.31 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 138.17 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 130.85 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 114.09 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 188.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 206.11 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:51 178.67 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 131.91 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 146.05 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 124.50 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 183.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 199.30 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 175.94 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 128.23 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 139.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 122.16 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 176.00 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 160.47 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 198.83 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 181.45 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 163.92 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 147.96 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 134.19 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 115.63 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 155.64 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 131.54 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 123.59 KB
fmhaSm100fKernel_QkvFp16OFp16H512PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 94.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.60 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 129.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 114.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 124.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 117.82 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 110.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 139.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 90.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 99.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 120.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 120.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 113.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 122.28 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 110.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 109.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 130.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 89.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 112.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 86.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 121.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 107.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 115.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 101.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 128.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 113.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 111.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 93.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 110.51 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 96.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 130.56 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 116.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 95.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 78.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 105.20 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 115.10 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 97.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 90.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 126.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 109.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 119.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 114.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 105.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 136.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 97.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 118.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 117.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 108.75 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 117.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 107.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 104.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 127.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 95.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 110.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 84.69 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 118.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 104.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 111.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 96.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 124.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 109.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 91.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 105.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 92.02 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.01 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 127.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 113.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 94.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 77.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 103.88 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 86.35 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 112.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 94.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 89.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.11 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 124.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 117.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 113.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 103.70 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 135.18 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 96.97 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 116.71 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.11 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 115.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 106.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 115.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 105.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 102.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 125.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 94.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 109.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 83.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 117.39 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 103.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 108.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 94.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 122.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 106.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 89.54 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 103.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 89.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.47 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 126.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 112.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 93.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 102.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 110.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 93.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 88.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 130.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 138.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 121.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 129.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 105.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 126.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 113.03 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 127.34 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 135.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 118.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 126.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 115.92 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 102.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 123.50 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 110.19 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 125.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 134.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 116.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 124.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 114.38 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 101.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 121.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvCustomP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 108.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.13 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 136.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 118.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 129.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 124.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 114.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 145.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 94.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 103.72 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 126.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 90.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 127.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 117.79 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 127.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 116.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 113.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 135.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.29 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 101.66 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 119.08 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 89.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 129.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 114.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 119.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 104.99 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 133.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 118.67 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 116.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 99.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 113.89 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 99.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 136.46 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 130.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 121.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 100.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 83.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 110.73 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 92.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 120.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 102.80 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 95.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 78.32 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 134.17 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 114.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 125.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 121.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 110.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 142.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 102.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 124.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 89.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 125.31 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 113.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 123.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 113.86 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 109.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 133.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 92.00 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 100.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 116.77 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.48 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 126.64 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 111.58 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 114.90 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 100.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 129.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 114.42 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 114.43 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 96.25 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 109.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 95.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 133.74 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.87 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 119.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 98.62 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 81.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 109.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 91.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 118.14 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 99.96 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 93.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 132.63 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 112.27 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 122.95 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 119.91 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 108.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 141.23 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 92.30 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 101.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 122.93 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.83 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 123.78 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 111.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 120.98 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 112.33 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.04 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 131.76 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 91.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 99.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 115.36 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.65 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.94 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 125.09 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.85 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 110.16 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 112.59 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 98.57 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 127.12 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 112.26 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 112.40 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 94.84 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 107.05 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.41 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.44 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 132.22 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.45 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 117.68 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 97.52 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 80.81 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 108.07 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ32Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 90.49 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128PersistentKeepsAbForGen.cubin 16-Apr-2026 13:50 116.24 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ64Kv128StaticKeepsAbForGen.cubin 16-Apr-2026 13:50 98.55 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 92.21 KB
fmhaSm100fKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.46 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.13 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 97.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 75.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.34 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 70.33 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 62.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 60.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 69.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.26 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.45 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 60.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 66.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 58.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 68.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 68.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 97.41 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 66.33 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 58.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 64.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 56.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 79.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 75.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 78.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 76.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 64.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 73.31 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 61.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 77.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.01 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 73.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 61.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 70.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 59.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 104.59 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 71.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 59.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 57.75 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 97.53 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 101.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 85.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 101.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 98.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 99.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 97.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.73 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 95.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 92.83 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 83.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 94.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 91.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 90.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.29 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 86.58 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 79.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 84.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 76.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 90.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 86.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.65 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 101.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.67 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 82.58 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 79.95 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 104.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 101.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 102.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 99.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 98.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 87.25 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 95.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 85.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 98.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 94.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 96.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 97.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 92.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 79.88 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 89.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 77.50 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 93.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 89.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 91.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 88.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 88.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 75.38 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 85.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.21 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.91 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.52 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.61 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.64 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 143.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 139.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 140.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 137.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 138.81 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 126.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 135.17 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 123.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 130.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 126.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 127.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 124.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 123.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 113.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 119.55 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 111.62 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 124.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 120.44 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 121.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 118.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 116.50 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.62 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 112.92 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 105.24 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 147.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 142.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 143.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 140.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 142.44 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 125.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 138.28 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 123.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 133.94 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 129.69 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 130.99 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 127.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 130.04 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 115.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 126.03 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 112.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 127.82 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 123.44 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 124.87 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 121.74 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 123.16 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 107.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 119.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H512PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 104.35 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 91.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 83.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 89.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 80.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 87.10 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 78.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 91.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 83.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 60.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 66.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 58.77 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 68.85 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 69.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 68.02 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 89.72 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 80.56 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 66.48 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 58.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 64.20 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 56.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 69.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 67.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 68.18 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 66.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 87.22 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 78.43 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 64.23 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 57.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 62.08 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvDenseP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 55.15 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 75.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.60 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 74.80 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 98.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.32 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 74.09 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 61.79 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 71.06 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 59.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 73.63 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.47 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 70.90 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 96.40 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 71.71 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 59.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.68 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 57.50 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 72.00 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvCgaVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 69.97 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 71.05 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64MultiCtasKvVarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 69.27 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 93.98 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 85.07 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 68.84 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ16Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 57.66 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128PersistentSwapsAbForGen.cubin 16-Apr-2026 13:50 65.93 KB
fmhaSm103aKernel_QE4m3KvE2m1OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ8Kv128StaticSwapsAbForGen.cubin 16-Apr-2026 13:50 55.50 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.34 KB
fmhaSm103aKernel_QkBfloat16VE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.63 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.17 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.19 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.57 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.14 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.31 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.40 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.77 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.06 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.97 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.71 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 134.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 151.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 138.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 132.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.89 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.62 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.78 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.31 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.43 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.17 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.47 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.28 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 144.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 131.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 133.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.39 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.49 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 155.77 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 137.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 159.22 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 140.60 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 144.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.37 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.75 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.58 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.11 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.90 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.11 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.21 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 166.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 155.01 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 169.33 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 157.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 144.12 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 131.48 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.09 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 164.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 157.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 167.52 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 160.51 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 144.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 132.03 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.08 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 134.98 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.05 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.65 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.46 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.59 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.71 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.61 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.92 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 108.02 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 94.32 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.84 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.45 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.18 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.99 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.16 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.38 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.68 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.27 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.54 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.13 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.56 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.73 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.00 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.23 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.79 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.30 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.80 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.70 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.81 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.41 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.26 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.10 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.72 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.07 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.36 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.20 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.96 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.85 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.82 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.66 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.42 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.95 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.83 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.64 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.44 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.61 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.35 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.50 KB
fmhaSm103aKernel_QkvBfloat16OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.10 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.97 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 120.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 107.43 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 127.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 113.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 109.49 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.73 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.78 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.67 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.37 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.21 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.38 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.03 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.97 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.82 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.01 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 107.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 127.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 113.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 109.49 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.91 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 103.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.93 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 111.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.18 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.42 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.63 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.14 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.93 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.87 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.01 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.14 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.85 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.46 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.47 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.58 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 103.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.74 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.93 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.17 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H256SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.69 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.94 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.29 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.22 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.06 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.36 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.51 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.08 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.07 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.40 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.20 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.13 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.45 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.93 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.80 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.77 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.76 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.35 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.86 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 119.50 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 94.71 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.96 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 100.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.41 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 96.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.88 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.04 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.75 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.90 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.23 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.27 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.25 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.33 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.68 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.12 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.63 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.48 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.34 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.55 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.67 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.70 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.66 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.01 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.15 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.16 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.02 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.68 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.44 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.10 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.19 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.59 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.56 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.05 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.24 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 119.62 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 94.71 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 100.79 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.53 KB
fmhaSm103aKernel_QkvE4m3OBfloat16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 96.65 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.98 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.64 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.57 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.52 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.09 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.39 KB
fmhaSm103aKernel_QkvE4m3OBfloat16HQk192HV128SeparateQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.55 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 97.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 111.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 123.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 109.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 129.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 115.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 124.94 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 111.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.11 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.03 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.74 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.32 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 143.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.61 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.07 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.20 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 111.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 123.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 109.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 129.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 115.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 125.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 111.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.32 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 101.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.76 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 91.31 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 101.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.96 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 91.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.40 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.30 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.69 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.28 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.62 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.34 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.88 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.83 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.76 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.79 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.06 KB
fmhaSm103aKernel_QkvE4m3OE2m1H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.67 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.24 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.19 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.63 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.90 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.12 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.53 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.75 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.57 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.64 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.52 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.23 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.92 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.70 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.35 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 130.03 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.25 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.59 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.05 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.93 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.72 KB
fmhaSm103aKernel_QkvE4m3OE2m1H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.56 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.69 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.16 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.68 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.33 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 109.51 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 96.14 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 115.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 102.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 111.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 98.08 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.50 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.10 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.28 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.73 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.49 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.67 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.95 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.64 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.97 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.84 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.47 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.43 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.31 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.54 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.77 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.81 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.86 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.44 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.98 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.37 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.34 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.87 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.45 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.91 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.02 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.38 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.48 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.80 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.39 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.67 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.26 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.82 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.85 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 109.63 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 96.27 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 115.66 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 102.22 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 111.60 KB
fmhaSm103aKernel_QkvE4m3OE2m1H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 98.08 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.58 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.80 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 133.83 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 120.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 140.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 126.87 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 135.81 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 122.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.57 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.60 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.89 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.65 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.43 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.10 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.83 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.97 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.36 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.00 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 131.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 134.25 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.17 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.24 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 131.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.71 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 143.54 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 133.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 120.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 140.14 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 126.99 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 135.94 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 122.57 KB
fmhaSm103aKernel_QkvE4m3OE4m3H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 91.99 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.33 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.74 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.26 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.93 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.74 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.44 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.91 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.59 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.09 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.83 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.76 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.86 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.75 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.57 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.06 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.99 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.40 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.14 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.02 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.84 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.22 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 130.67 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 150.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 134.11 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.69 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.11 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.81 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.10 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.64 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 131.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 151.42 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 134.19 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.95 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.38 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.93 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.07 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.05 KB
fmhaSm103aKernel_QkvE4m3OE4m3H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.21 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.53 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.20 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 92.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 102.86 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 113.18 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 99.33 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 119.37 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 105.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 115.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 101.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.34 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.88 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.33 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 112.75 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.77 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.23 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.15 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.51 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.48 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.06 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.04 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.68 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.28 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.63 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.50 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.78 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.62 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.82 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.70 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.81 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.01 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.17 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.16 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.29 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.06 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.86 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.55 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.31 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.98 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.06 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.47 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.52 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.56 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.66 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.92 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.12 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 113.30 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 99.45 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 119.37 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 105.41 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 115.27 KB
fmhaSm103aKernel_QkvE4m3OE4m3H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 101.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.48 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.22 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.08 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 107.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 127.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 113.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.87 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 109.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 97.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.75 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.63 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.49 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.42 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.72 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.76 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.36 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.82 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.87 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.27 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.47 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.15 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 121.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 107.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 127.20 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 113.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.99 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 109.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 97.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.82 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 87.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 104.37 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 91.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 100.96 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 88.19 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 105.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.94 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 111.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 98.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.67 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.36 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.34 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.83 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.81 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.73 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.97 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.82 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.05 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.07 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.68 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.89 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.47 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 119.19 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.10 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.45 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.72 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.02 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.11 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.89 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.53 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.19 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.12 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.47 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 99.36 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.62 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.78 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.89 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.46 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.17 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 127.59 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.61 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.39 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.60 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.30 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.18 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.03 KB
fmhaSm103aKernel_QkvE4m3OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 103.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.11 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 103.73 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 120.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 94.71 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 100.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.16 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 96.65 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 89.93 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.77 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.51 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.66 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.98 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.03 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.24 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.01 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.09 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.92 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.41 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.38 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.88 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.52 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 107.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.54 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.23 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.43 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.58 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.89 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.91 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 121.04 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.90 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.74 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.44 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.20 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.86 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.95 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.35 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.40 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.32 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.55 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.80 KB
fmhaSm103aKernel_QkvE4m3OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.00 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1PersistentContext.cubin 16-Apr-2026 13:50 120.51 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK16SageV1StaticContext.cubin 16-Apr-2026 13:50 94.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.84 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK1SageV1StaticContext.cubin 16-Apr-2026 13:50 100.79 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1PersistentContext.cubin 16-Apr-2026 13:50 122.29 KB
fmhaSm103aKernel_QkvE4m3OFp16H64SeparateQkvDenseVarSeqQ128Kv128SageQ1SageK4SageV1StaticContext.cubin 16-Apr-2026 13:50 96.65 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 113.76 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 100.77 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.50 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.30 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.79 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.33 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.71 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.33 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.41 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.16 KB
fmhaSm103aKernel_QkvFp16OFp16H128PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.40 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.27 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.45 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.12 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.35 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.53 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.91 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.20 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.58 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.65 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.13 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.35 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 137.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.98 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.65 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 128.96 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.23 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.31 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.88 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 123.17 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.59 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.46 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.84 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 136.00 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 152.39 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 139.41 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 143.80 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.95 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 147.34 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 133.49 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 140.03 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 126.20 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 143.57 KB
fmhaSm103aKernel_QkvFp16OFp16H128PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.73 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 114.43 KB
fmhaSm103aKernel_QkvFp16OFp16H128SeparateQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.91 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 107.26 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 90.97 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 94.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 106.96 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 93.98 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 110.06 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 96.93 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 117.68 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 101.56 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.68 KB
fmhaSm103aKernel_QkvFp16OFp16H256PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 104.84 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.19 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.39 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 149.40 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 130.30 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.24 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 111.50 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.60 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 114.41 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.05 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.55 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.18 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.59 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 146.11 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 132.15 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.98 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 135.09 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.14 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.33 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.43 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.15 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.52 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.30 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 125.62 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.12 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 156.90 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 138.70 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 160.36 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 141.74 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 142.05 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.79 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.44 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.83 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.50 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.97 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.57 KB
fmhaSm103aKernel_QkvFp16OFp16H256PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.89 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.13 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.24 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.96 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.24 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.52 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 105.27 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 118.80 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.34 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 167.37 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 156.14 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 170.59 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 158.94 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.13 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 129.84 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.36 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 132.62 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 131.61 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 118.44 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.97 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.23 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 165.31 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 158.82 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 168.65 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 161.64 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 145.37 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 133.29 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 148.22 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 136.12 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 132.19 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 121.91 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 135.60 KB
fmhaSm103aKernel_QkvFp16OFp16H512HVPerCta256PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 124.73 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.35 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 115.70 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 102.01 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 109.66 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 95.96 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 116.30 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvDenseVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 103.05 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 120.02 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 106.09 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.27 KB
fmhaSm103aKernel_QkvFp16OFp16H64PackedQkvSlidingOrChunkedCausalVarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 113.07 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.92 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.02 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.47 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.41 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.74 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.48 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.62 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 122.90 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 108.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 128.94 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 115.20 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 126.95 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 112.87 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 133.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.38 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 124.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 110.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 130.91 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 117.58 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 123.45 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 109.05 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 129.34 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvDenseP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 116.18 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 138.36 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 122.99 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 144.31 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP16VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 130.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 136.00 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 120.84 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 141.33 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP32VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 127.22 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 134.46 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 119.30 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128PersistentContext.cubin 16-Apr-2026 13:50 139.78 KB
fmhaSm103aKernel_QkvFp16OFp16H64PagedKvSlidingOrChunkedCausalP64VarSeqSkipsSoftmaxQ128Kv128StaticContext.cubin 16-Apr-2026 13:50 125.80 KB
LICENSE 16-Apr-2026 13:51 11.09 KB
Artifactory/7.117.25 Server