bf16xint16_gemm operator: add --transpose option #2466
Closed
Conversation
This file contains 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
davidberard98
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
|
@davidberard98 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
davidberard98
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
Summary: `--transpose` will make this benchmark test a int16 x bf16 mm instead of a bf16 x int16. This matters for H100, because the wgmma instruction can take registers only on the LHS. So int16 x bf16 is probably the easier one to support efficiently. Pull Request resolved: pytorch#2466 Test Plan: In OSS: ran `python run_benchmark.py triton --op bf16xint16_gemm --transpose` Internally, ran `buck2 run mode/opt //pytorch/benchmark:triton -- --op bf16xint16_gemm --transpose` Internally, we run into the issue fixed by triton-lang/triton#4695; but otherwise, they both run. Differential Revision: D63294109 Pulled By: davidberard98
|
This pull request was exported from Phabricator. Differential Revision: D63294109 |
davidberard98
force-pushed
the
bf16xint16-transpose
branch
from
dad7188
to
edeb11a
Compare
davidberard98
had a problem deploying
to
docker-s3-upload
GitHub Actions
Failure
davidberard98
had a problem deploying
to
docker-s3-upload
GitHub Actions
Failure
|
@davidberard98 merged this pull request in 0ab0e47. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
--transposewill make this benchmark test a int16 x bf16 mm instead of a bf16 x int16.This matters for H100, because the wgmma instruction can take registers only on the LHS. So int16 x bf16 is probably the easier one to support efficiently.