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 illegal memory address when skipping -1 m indices (#113)

Browse Source 
Co-authored-by: Shixian Cui <shixian@amazon.com>
This commit is contained in:
shixianc 2025-06-15 19:44:31 -07:00 committed by GitHub
parent 8dfa329827
commit 0c88cd0139
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

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

@ -116,7 +116,7 @@ struct Scheduler {
if constexpr (kGemmType == GemmType::Normal) {
return block_idx * block_size;
} else if constexpr (kGemmType == GemmType::GroupedContiguous) {
auto offset = kIgnoreGroupedForGroupedContiguous ? 0 : __ldg(grouped_layout + m_block_idx * BLOCK_M);
auto offset = kIgnoreGroupedForGroupedContiguous ? 0 : max(0, __ldg(grouped_layout + m_block_idx * BLOCK_M));
return offset * shape_dim + block_idx * block_size;
} else if constexpr (kGemmType == GemmType::GroupedMasked) {
return curr_group_idx * shape_dim + block_idx * block_size;