Tune threads per block for mscclpp executor (#345)

This commit is contained in:
Binyang Li
2024-09-18 17:21:47 -07:00
committed by GitHub
parent 0c7311e83f
commit b30bb260e3
12 changed files with 43 additions and 46 deletions

View File

@@ -630,7 +630,6 @@ def test_executor(mpi_group: MpiGroup, filename: str):
sendbuf.nbytes,
sendbuf.nbytes,
DataType.float16,
512,
execution_plan,
stream.ptr,
)