Yi DING 820ba182a0 Fix unmatched K size of WarpGemmMfmaBf16Bf16F32M16N16K32TransposedCDistribution on gfx950 (#2393) ...

[ROCm/composable_kernel commit: c5d9181e1b]

2025-06-24 16:35:54 -07:00

..

add_rmsnorm2d_rdquant

…

batched_transpose

…

common

…

elementwise

…

epilogue

…

flatmm

…

fmha

[CK_TILE] FMHA Support hdim_v to as a Multiple of 32 (#2114)

2025-06-24 01:33:31 +08:00

fused_moe

[CK_TILE] Fix compilation errors introduced in #2320, #2219 and #2214 (#2388)

2025-06-23 12:29:15 +08:00

gemm

Fix unmatched K size of WarpGemmMfmaBf16Bf16F32M16N16K32TransposedCDistribution on gfx950 (#2393)

2025-06-24 16:35:54 -07:00

grouped_convolution

…

image_to_column

…

layernorm2d

…

norm_reduce

[CK_TILE] Fix compilation errors introduced in #2320, #2219 and #2214 (#2388)

2025-06-23 12:29:15 +08:00

permute

…

reduce/block

…

rmsnorm2d

…

smoothquant

…

softmax/block

…

topk/block

…

topk_softmax

…

add_rmsnorm2d_rdquant.hpp

…

batched_transpose.hpp

…

common.hpp

…

elementwise.hpp

…

epilogue.hpp

…

flatmm.hpp

…

fmha.hpp

…

fused_moe.hpp

…

gemm.hpp

…

grouped_convolution.hpp

…

image_to_column.hpp

…

layernorm2d.hpp

…

norm_reduce.hpp

…

permute.hpp

…

reduce.hpp

…

rmsnorm2d.hpp

…

smoothquant.hpp

…

softmax.hpp

…

topk_softmax.hpp

…

topk.hpp

…