Updates for 3.1 (#932)

2023-04-29 09:34:27 -04:00
parent 6f8596ce3f
commit 7c04f95415
51 changed files with 1796 additions and 328 deletions
@@ -150,7 +150,7 @@ template<
  int Stages_,
  class ClusterShape_ = Shape<_1,_1,_1>,
  class KernelSchedule = KernelTmaWarpSpecialized,
-  int PipelineAsyncMmaStages_ = 1
+  int PipelineAsyncMmaStages_ = 0
 >
 struct MainloopSm90TmaGmmaRmemAWarpSpecialized {
  constexpr static int Stages = Stages_;