frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CUDA][FP8] Skip rowwise scaling test on sm89 (#135718)

Browse Source 
Same reason as #https://github.com/pytorch/pytorch/pull/133612, rowwise scaling implementation is sm90+ specific (e.g., uses TMA)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/135718
Approved by: https://github.com/Skylion007
This commit is contained in:
eqy
2024-09-13 15:07:18 +00:00
committed by PyTorch MergeBot
parent ba6e0f31ab
commit 2519e5a8de
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
test/test_matmul_cuda.py
View File

@ -560,6 +560,7 @@ class TestFP8MatmulCuda(TestCase):
self.assertEqual(out_fp8, out_fp8_s)
@unittest.skipIf(not PLATFORM_SUPPORTS_FP8 or IS_WINDOWS, f8_msg)
@unittest.skipIf(not SM90OrLater, "rowwise implementation is currently sm90 specific")
@skipIfRocm()
@parametrize("use_fast_accum", [True, False])
def test_float8_rowwise_scaling_sanity(self, device, use_fast_accum: bool) -> None: