[PyTorch] Port ExecuTorch bfdot improvement back to ATen BlasKernel, Try #2 #137377

swolchok · 2024-10-04T22:29:40Z

Stack from ghstack (oldest at bottom):

ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 .

Differential Revision: D63923166

…Try #2 ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

pytorch-bot · 2024-10-04T22:29:44Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/137377

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit fcb5288 with merge base de4c2a3 ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

facebook-github-bot · 2024-10-04T22:29:49Z

This pull request was exported from Phabricator. Differential Revision: D63923166

…Try #2 ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) ghstack-source-id: 246411194 Pull Request resolved: #137377

swolchok · 2024-10-04T22:30:42Z

@pytorchbot label "ciflow/linux-aarch64"

…lasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-07T15:06:23Z

This pull request was exported from Phabricator. Differential Revision: D63923166

…lasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-07T16:43:41Z

This pull request was exported from Phabricator. Differential Revision: D63923166

…Try #2 Pull Request resolved: #137377 ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . ghstack-source-id: 246616406 Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/)

albanD · 2024-10-07T16:51:43Z

What's the difference with the previous attempt?

swolchok · 2024-10-08T02:10:13Z

aten/src/ATen/native/BlasKernel.cpp

+  return reduce(sum);
 }

+// NOTE: The first attempt at landing BFDOT support with


@albanD here is the difference

Ho thanks for the pointer!
Should we undef DOT_WITH_FP32_ARITH_TAIL_AFTER_MAIN_LOOP_BODY ?

this is a .cpp file, so the macro isn't going to leak anywhere and it's not particularly necessary, but sure I can do that.

…lasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-08T17:19:02Z

This pull request was exported from Phabricator. Differential Revision: D63923166

swolchok · 2024-10-08T18:33:59Z

will fix https://hud.pytorch.org/pr/pytorch/pytorch/137377#31249893287

… back to ATen BlasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-08T20:18:13Z

This pull request was exported from Phabricator. Differential Revision: D63923166

…Ten BlasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-08T23:49:29Z

This pull request was exported from Phabricator. Differential Revision: D63923166

…Try #2 Pull Request resolved: #137377 ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . ghstack-source-id: 246956192 Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/)

swolchok · 2024-10-09T22:37:26Z

CI is 100% green; please review

…lasKernel, Try #2" ExecuTorch's fork of BlasKernel.cpp grew bfdot support, complete with demonstration that it helps. Port it back to PyTorch. First attempt was #136331 . Differential Revision: [D63923166](https://our.internmc.facebook.com/intern/diff/D63923166/) [ghstack-poisoned]

facebook-github-bot · 2024-10-10T04:22:11Z

This pull request was exported from Phabricator. Differential Revision: D63923166

malfet · 2024-10-10T05:20:04Z

If CI is green, than sure, LGTM

swolchok · 2024-10-10T19:36:39Z

@pytorchbot merge

pytorchmergebot · 2024-10-10T19:38:59Z

Merge started

Your 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

Advanced Debugging

Check the merge workflow status
here

facebook-github-bot added the fb-exported label Oct 4, 2024

swolchok mentioned this pull request Oct 4, 2024

[PyTorch] Port ExecuTorch bfdot improvement back to ATen BlasKernel #136331

Closed

pytorch-bot bot added the ciflow/linux-aarch64 linux aarch64 CI workflow label Oct 4, 2024

swolchok added release notes: performance_as_product topic: performance topic category labels Oct 4, 2024

swolchok requested review from albanD and malfet and removed request for albanD October 4, 2024 22:31

swolchok mentioned this pull request Oct 7, 2024

[PyTorch] Use 128-bit vectors for ARM64 #137426

Closed

swolchok commented Oct 8, 2024

View reviewed changes

swolchok added the ciflow/trunk Trigger trunk jobs on your pull request label Oct 8, 2024

swolchok mentioned this pull request Oct 9, 2024

[PyTorch] Move NEON VecConvert specialization from vec256_convert to vec128_convert #137661

Closed

malfet approved these changes Oct 10, 2024

View reviewed changes

This was referenced Oct 10, 2024

[PyTorch] Check defined(__aarch64__) && !defined(CPU_CAPABILITY_SVE256) instead of defined(CPU_CAPABILITY_NEON) #137722

Closed

[PyTorch] add NEON half2float fmadd/fmsub #137723

Closed

pytorchmergebot added the merging label Oct 10, 2024

pytorchmergebot added the Merged label Oct 10, 2024

pytorchmergebot closed this in 9c12198 Oct 10, 2024

pytorchmergebot removed the merging label Oct 10, 2024

github-actions bot deleted the gh/swolchok/649/head branch November 10, 2024 02:07

[PyTorch] Port ExecuTorch bfdot improvement back to ATen BlasKernel, Try #2 #137377

[PyTorch] Port ExecuTorch bfdot improvement back to ATen BlasKernel, Try #2 #137377

Uh oh!

Conversation

swolchok commented Oct 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Oct 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/137377

✅ No Failures

Uh oh!

facebook-github-bot commented Oct 4, 2024

Uh oh!

swolchok commented Oct 4, 2024

Uh oh!

facebook-github-bot commented Oct 7, 2024

Uh oh!

facebook-github-bot commented Oct 7, 2024

Uh oh!

albanD commented Oct 7, 2024

Uh oh!

swolchok Oct 8, 2024

Choose a reason for hiding this comment

Uh oh!

albanD Oct 8, 2024

Choose a reason for hiding this comment

Uh oh!

swolchok Oct 8, 2024

Choose a reason for hiding this comment

Uh oh!

swolchok Oct 8, 2024

Choose a reason for hiding this comment

Uh oh!

facebook-github-bot commented Oct 8, 2024

Uh oh!

swolchok commented Oct 8, 2024

Uh oh!

facebook-github-bot commented Oct 8, 2024

Uh oh!

facebook-github-bot commented Oct 8, 2024

Uh oh!

swolchok commented Oct 9, 2024

Uh oh!

facebook-github-bot commented Oct 10, 2024

Uh oh!

malfet commented Oct 10, 2024

Uh oh!

swolchok commented Oct 10, 2024

Uh oh!

pytorchmergebot commented Oct 10, 2024

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

swolchok commented Oct 4, 2024 •

edited

Loading

pytorch-bot bot commented Oct 4, 2024 •

edited

Loading