ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-19 14:29:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix MX bpreshuffle gemm B grid descriptor dimension error. (#3170)

Browse Source
This commit is contained in:
Gino Lu
2025-11-07 11:42:39 +08:00
committed by GitHub
parent d04eba4ae3
commit e31a7a4f29
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_v3_mx_bpreshuffle.hpp
View File

@@ -429,8 +429,8 @@ struct GridwiseGemmMX_xdl_cshuffle_v3_bpreshuffle
constexpr index_t MWave = MPerBlock / (MXdlPerWave * MPerXdl);
constexpr index_t WaveSize = BlockSize / (MWave * NWave);
constexpr index_t NkSwizzleNumber = Number<WaveSize * KPack>{};
return make_naive_tensor_descriptor_packed(
make_tuple(N0 / NWave / NXdlPack, NWave, NXdlPack, K0, NkSwizzleNumber));
return make_naive_tensor_descriptor_packed(make_tuple(
math::integer_divide_ceil(N0, NWave * NXdlPack), NWave, NXdlPack, K0, NkSwizzleNumber));
}
__host__ __device__ static auto MakeBGridDescriptor_BK0_N_BK1(