Enable tf32 precision acceleration for h100 #4767
Open
+20
−0
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
added 2 commits
August 24, 2025 14:40
Summary: X-link: facebookresearch/FBGEMM#1788 For H100, add new option to persistent triton fp8 gemm to boost perf for CMF 500x gemm shapes: - M=512, N=1024, K=19712, boost flops from 504 to 559, by 11%. - M=512, N=1024, K=171712, boost flops from 437 to 481, by 10%. E2E - CMF 500x QPS was [25077.38](https://www.internalfb.com/intern/paste/P1915458454/), now it is [25319.47](https://www.internalfb.com/intern/paste/P1916364403/). Differential Revision: D80881599
Summary: As indicated by previous MTS [benchmark](https://www.internalfb.com/intern/everpaste/?handle=GEHWGSBdCZC6Q5QFALLQJHKASQUnbsIXAAAB&phabricator_paste_number=1916364403): ``` TensorFloat32 tensor cores for float32 matrix multiplication available but not enabled. Consider setting `torch.set_float32_matmul_precision('high')` for better performance. ``` By this diff, enable fp32 high precision. E2E, the diff pushes CMF500x QPS from [25319.47](https://www.internalfb.com/intern/everpaste/?handle=GEHWGSBdCZC6Q5QFALLQJHKASQUnbsIXAAAB&phabricator_paste_number=1916364403) to [25737.61](https://www.internalfb.com/intern/everpaste/?handle=GAsZ7R9ASEeoe9sIABHwORv6SG89bsIXAAAB&phabricator_paste_number=1916396963) Differential Revision: D80908603
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
This pull request was exported from Phabricator. Differential Revision: D80908603 |
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Outdated suggestions cannot be applied.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.
Summary:
As indicated by previous MTS benchmark:
By this diff, enable fp32 high precision. E2E, the diff pushes CMF500x QPS from 25319.47 to 25737.61
Differential Revision: D80908603