[Kernel] Fix CUTLASS 3.x custom broadcast load epilogue (#5516)

2024-06-14 12:30:15 -04:00
parent d47af2bc02
commit 703475f6c2
2 changed files with 2 additions and 4 deletions
--- a/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp
+++ b/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp
@@ -153,7 +153,7 @@ struct Sm90RowOrScalarBroadcast {

    CUTLASS_DEVICE void
    begin(uint64_t* full_mbarrier_ptr, int load_iteration, bool issue_tma_load) {
-      if (params.ptr_row == nullptr) {
+      if (!params.row_broadcast) {
        return;
      }