ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-07-09 00:27:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Bug fix.

Browse Source
This commit is contained in:
Ville Pietilä
2025-08-27 14:30:13 +00:00
parent f1a7cfba26
commit b82e68c45f
1 changed files with 5 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_v3.hpp
View File

@@ -1375,6 +1375,7 @@ struct GridwiseGemm_xdl_cshuffle_v3
{
#if defined(__gfx950__)
return
!DoElementwiseBeforeCShuffle &&
std::is_same_v<CShuffleDataType, ck::bhalf_t> &&
std::is_same_v<AccDataType, float>;
#else
@@ -1651,9 +1652,7 @@ struct GridwiseGemm_xdl_cshuffle_v3
CShuffleDataType,
decltype(c_thread_desc_m0_n0_m1_n1_m2_m3_m4_n2),
decltype(c_block_desc_m0_n0_m1_n1_m2_m3_m4_n2),
conditional_t<DoElementwiseBeforeCShuffle,
CElementwiseOperation,
tensor_operation::element_wise::PassThrough>,
tensor_operation::element_wise::PassThrough,
Sequence<CShuffleMXdlPerWavePerShuffle,
CShuffleNXdlPerWavePerShuffle,
I1,
@@ -2117,7 +2116,9 @@ struct GridwiseGemm_xdl_cshuffle_v3
CShuffleDataType,
decltype(c_thread_desc_m0_n0_m1_n1_m2_m3_m4_n2),
decltype(c_block_desc_m0_n0_m1_n1_m2_m3_m4_n2),
ck::tensor_operation::element_wise::PassThrough,
conditional_t<DoElementwiseBeforeCShuffle,
CElementwiseOperation,
tensor_operation::element_wise::PassThrough>,
Sequence<CShuffleMXdlPerWavePerShuffle,
CShuffleNXdlPerWavePerShuffle,
I1,