-
Notifications
You must be signed in to change notification settings - Fork 749
make triton kernel usage user controlable #16030
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16030
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New Failures, 1 Unrelated FailureAs of commit a84c6ea with merge base a93f59e ( NEW FAILURES - The following jobs have failed:
BROKEN TRUNK - The following job failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
@Gasoonjia has exported this pull request. If you are a Meta employee, you can view the originating Diff in D88096054. |
larryliu0820
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Review automatically exported from Phabricator review in Meta.
This PR needs a
|
Summary: We observed perf regression on some models after using triton sdpa kernel, compare with decomposed operator. This diff makes triton kernel controalable by user to prevent such perf regression. Reviewed By: larryliu0820 Differential Revision: D88096054
e43441d to
7f6f9eb
Compare
Summary: We observed perf regression on some models after using triton sdpa kernel, compare with decomposed operator. This diff makes triton kernel controalable by user to prevent such perf regression. Reviewed By: larryliu0820 Differential Revision: D88096054
7f6f9eb to
03e7403
Compare
Summary: We observed perf regression on some models after using triton sdpa kernel, compare with decomposed operator. This diff makes triton kernel controalable by user to prevent such perf regression. Reviewed By: larryliu0820 Differential Revision: D88096054
03e7403 to
a84c6ea
Compare
Summary: We observed perf regression on some models after using triton sdpa kernel, compare with decomposed operator. This diff makes triton kernel controalable by user to prevent such perf regression.
Reviewed By: larryliu0820
Differential Revision: D88096054