flash: Fix grid size to hw cluster size

Verified fast config, minus the barrier stall at the end.
2024-09-09 15:43:31 -07:00
parent 829af5d429
commit d31c8ffd7d
2 changed files with 3 additions and 7 deletions
--- a/tests/regression/sgemm_tcore/sgemm_impl.hpp
+++ b/tests/regression/sgemm_tcore/sgemm_impl.hpp
@@ -538,6 +538,7 @@ wmma_store(const int thread_in_warp, const int warp_col, const int warp_row,

 __attribute__((convergent)) inline void
 threadblock_barrier(const uint32_t barrier_id, const uint32_t count) {
+  asm volatile("" ::: "memory");
  vx_fence();
  vx_barrier(barrier_id, count);
 }