Insert barrier after loading work items to LDS (#1551)

[ROCm/rccl commit: baaa2ac64d]
2025-02-18 10:17:27 -08:00
parent 35987b9170
commit 7eff149ceb
1 changed files with 1 additions and 1 deletions
@@ -560,8 +560,8 @@ __device__ __forceinline__ void ncclKernelMain(struct ncclDevKernelArgs const* a
      default:
        break;
    }
-    __synclds();
    loadWorkBatchToShmem(tid%WARP_SIZE, tn, args, batchIx);
+    __synclds();

    // Check whether the last operation was aborted and make sure all threads exit
    bool aborted = false;