DeepSeek/DeepGEMM
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepGEMM synced 2025-06-26 23:15:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix TMA store bugs and code format

Browse Source
This commit is contained in:
Chenggang Zhao
2025-02-27 17:57:21 +08:00
parent b05ed2f017
commit ca13ce0fab
2 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
deep_gemm/include/deep_gemm/fp8_gemm.cuh
View File

@@ -406,7 +406,8 @@ public:
template <typename T>
static CUtensorMap make_2d_tma_d_desc(T* global_address, uint32_t shape_m) {
return make_2d_tma_desc(global_address, Layout::RowMajor,
shape_m * (kGemmType == GemmType::GroupedMasked ? kNumGroups : 1), SHAPE_N, BLOCK_M, BLOCK_N,
shape_m * (kGemmType == GemmType::GroupedMasked ? kNumGroups : 1), SHAPE_N,
min(BLOCK_M, shape_m), BLOCK_N,
CUtensorMapSwizzle::CU_TENSOR_MAP_SWIZZLE_NONE);
}