DeepSeek/DeepGEMM
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepGEMM synced 2025-05-05 21:35:26 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Minor update

Browse Source
This commit is contained in:
Chenggang Zhao 2025-04-23 14:14:49 +08:00
parent d7c068d467
commit 24517316af
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/fp8_gemm.cuh
View File

@ -276,7 +276,7 @@ fp8_gemm_kernel(__nv_bfloat16* gmem_d, float* scales_b, int* grouped_layout,
// Empty barrier arrival // Empty barrier arrival
auto empty_barrier_arrive = [&](int s) { auto empty_barrier_arrive = [&](int s) {
if (kNumTMAMulticast == 1) { if constexpr (kNumTMAMulticast == 1) {
lane_idx == 0 ? empty_barriers[s]->arrive() : void(); lane_idx == 0 ? empty_barriers[s]->arrive() : void();
} else { } else {
auto target_cta = scheduler.is_peer_cta_alive ? lane_idx : cute::block_rank_in_cluster(); auto target_cta = scheduler.is_peer_cta_alive ? lane_idx : cute::block_rank_in_cluster();