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

small typo

Browse Source
This commit is contained in:
yukuai 2025-05-20 16:20:50 +08:00
parent db1f0b5a1c
commit dc89674f47
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
deep_gemm/include/deep_gemm/fp8_gemm.cuh
View File

@ -271,7 +271,7 @@ fp8_gemm_kernel(float* scales_b, int* grouped_layout,
}
};
if (!scheduler.is_valid_m(math_wg_idx * WGMMA::M, m_block_idx)) {
if (!scheduler.is_m_valid(math_wg_idx * WGMMA::M, m_block_idx)) {
// Skip useless computation for unaligned Ms
launch_k_iterations([&](int k_iter, auto type, auto _) {
#pragma unroll

2
deep_gemm/include/deep_gemm/scheduler.cuh
View File

@ -48,7 +48,7 @@ struct Scheduler {
}
}
__device__ __forceinline__ bool is_valid_m(const uint32_t m_offset, const uint32_t& m_block_idx) const {
__device__ __forceinline__ bool is_m_valid(const uint32_t m_offset, const uint32_t& m_block_idx) const {
if constexpr (kGemmType == GemmType::Normal) {
return true;
} else if constexpr (kGemmType == GemmType::GroupedContiguous) {