[ROCm] Fix mx fp8 and fp4 code after scaling refactor changes. #163127
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
PR pytorch#151360 added mx fp8 and fp4 support on ROCm. However on recent upstream, scaling function in Blas.cpp along with test_matmul_cuda changes trigerred failures. This patch corrects is_blockwise_1x32_scaling function code and fixes minor bug in test_matmul_cuda. Testing result on gfx950 w/ ROCm7.0 PYTORCH_TEST_WITH_ROCM=1 python test/test_matmul_cuda.py -k test_blockwise -v Ran 452 tests in 22.698s OK passed 111 This is same as before. (when PR 151360 was merged) Signed-off-by: Jagadish Krishnamoorthy <jagadish.krishnamoorthy@amd.com>
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/163127
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 905b7f0 with merge base 9009c4d ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
jeffdaily
requested changes
Sep 17, 2025
jeffdaily
approved these changes
Sep 19, 2025
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Signed-off-by: Jagadish Krishnamoorthy <jagadish.krishnamoorthy@amd.com>
Merge failedReason: New commits were pushed while merging. Please rerun the merge command. Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 mandatory check(s) failed. The first few are: Dig deeper by viewing the failures on hud |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
mansiag05
pushed a commit
to mansiag05/pytorch
that referenced
this pull request
Sep 22, 2025
…ch#163127) PR pytorch#151360 added mx fp8 and fp4 support on ROCm. 1. However, on recent upstream, scaling function in Blas.cpp along with test_matmul_cuda changes triggered failures. This patch corrects is_blockwise_1x32_scaling function code. 2. Fixes the m, n, k dimensions for ROCm mx case. 3. Modify FP4E2M1FN_LARGEST_POW2 (largest power of 2 representable in `torch.float4_e2m1fn_x2`) to 2. This resulted in higher SQNR value for mx fp4 test. Testing result on gfx950 w/ ROCm7.0 PYTORCH_TEST_WITH_ROCM=1 python test/test_matmul_cuda.py -k test_blockwise -v Ran 452 tests in 22.698s OK passed 111 This is same as before. (when PR 151360 was merged) Pull Request resolved: pytorch#163127 Approved by: https://github.com/jeffdaily Co-authored-by: Jeff Daily <jeff.daily@amd.com>
cleonard530
pushed a commit
to cleonard530/pytorch
that referenced
this pull request
Sep 22, 2025
…ch#163127) PR pytorch#151360 added mx fp8 and fp4 support on ROCm. 1. However, on recent upstream, scaling function in Blas.cpp along with test_matmul_cuda changes triggered failures. This patch corrects is_blockwise_1x32_scaling function code. 2. Fixes the m, n, k dimensions for ROCm mx case. 3. Modify FP4E2M1FN_LARGEST_POW2 (largest power of 2 representable in `torch.float4_e2m1fn_x2`) to 2. This resulted in higher SQNR value for mx fp4 test. Testing result on gfx950 w/ ROCm7.0 PYTORCH_TEST_WITH_ROCM=1 python test/test_matmul_cuda.py -k test_blockwise -v Ran 452 tests in 22.698s OK passed 111 This is same as before. (when PR 151360 was merged) Pull Request resolved: pytorch#163127 Approved by: https://github.com/jeffdaily Co-authored-by: Jeff Daily <jeff.daily@amd.com>
dsashidh
pushed a commit
to dsashidh/pytorch
that referenced
this pull request
Sep 26, 2025
…ch#163127) PR pytorch#151360 added mx fp8 and fp4 support on ROCm. 1. However, on recent upstream, scaling function in Blas.cpp along with test_matmul_cuda changes triggered failures. This patch corrects is_blockwise_1x32_scaling function code. 2. Fixes the m, n, k dimensions for ROCm mx case. 3. Modify FP4E2M1FN_LARGEST_POW2 (largest power of 2 representable in `torch.float4_e2m1fn_x2`) to 2. This resulted in higher SQNR value for mx fp4 test. Testing result on gfx950 w/ ROCm7.0 PYTORCH_TEST_WITH_ROCM=1 python test/test_matmul_cuda.py -k test_blockwise -v Ran 452 tests in 22.698s OK passed 111 This is same as before. (when PR 151360 was merged) Pull Request resolved: pytorch#163127 Approved by: https://github.com/jeffdaily Co-authored-by: Jeff Daily <jeff.daily@amd.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR #151360 added mx fp8 and fp4 support on ROCm.
However, on recent upstream, scaling function in Blas.cpp along with test_matmul_cuda changes triggered failures.
This patch corrects is_blockwise_1x32_scaling function code.
Fixes the m, n, k dimensions for ROCm mx case.
Modify FP4E2M1FN_LARGEST_POW2 (largest power of 2 representable in
torch.float4_e2m1fn_x2) to 2.This resulted in higher SQNR value for mx fp4 test.
Testing result on gfx950 w/ ROCm7.0
PYTORCH_TEST_WITH_ROCM=1 python test/test_matmul_cuda.py -k test_blockwise -v Ran 452 tests in 22.698s
OK passed 111
This is same as before. (when PR 151360 was merged)
cc @jeffdaily @sunway513 @jithunnair-amd @pruthvistony @ROCmSupport @dllehr-amd @jataylo @hongxiayang @naromero77amd