Skip to content

[ScaledMM] Switch inductor triton lowerings to new tma persistent by default #142296

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.

Sign up for GitHub

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

Jump to bottom
Closed
drisspg wants to merge 3 commits into from
Closed

[ScaledMM] Switch inductor triton lowerings to new tma persistent by default #142296

drisspg wants to merge 3 commits into from

Conversation

@drisspg
Copy link
Contributor

@drisspg drisspg commented Dec 7, 2024 •
edited
Loading

@drisspg
Update
6a44fd6 
[ghstack-poisoned]
@drisspg drisspg mentioned this pull request Dec 7, 2024
Closed
@pytorch-bot
Copy link

pytorch-bot bot commented Dec 7, 2024 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/142296

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

✅ No Failures

As of commit 88d3147 with merge base db379ed (image):
💚 Looks good so far! There are no failures yet. 💚

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

@drisspg drisspg added the topic: not user facing topic category label Dec 7, 2024
@drisspg
Update
f6a994a 
[ghstack-poisoned]
@aakhundov
Copy link
Contributor

aakhundov commented Dec 9, 2024

@drisspg I'm planning to reuse the same flag for the general mm / addmm lowerings. Do we want the persistent+TMA templates to be enabled by default for those, too?

@drisspg
Copy link
Contributor Author

drisspg commented Dec 9, 2024

@aakhundov My end goal view is that this is pretty much a debug flag that is helpful for testing and if a user encounters some weird behavior it will be a good thing to ablate via Elias bisector tool. Ultimately I think that inductor's heuristic should be smart enough to pick the best lowering for problem shape x hardware. In practice this ends up being a big cross product sometimes wrong decision are made. SO tbh idk I guess it depends on how confident we are in the feature. Maybe a JK rollout test before globally flipping

@drisspg
Update
88d3147 
[ghstack-poisoned]
drisspg added a commit that referenced this pull request Dec 9, 2024
@drisspg
[ScaledMM] Switch inductor triton lowerings to new tma persistent by …
c049a0a 
…default

ghstack-source-id: fd9dcd4
Pull Request resolved: #142296
@github-actions
Copy link
Contributor

github-actions bot commented Feb 7, 2025

Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as Stale.
Feel free to remove the Stale label if you feel this was a mistake.
If you are unable to remove the Stale label please contact a maintainer in order to do so.
If you want the bot to never mark this PR stale again, add the no-stale label.
Stale pull requests will automatically be closed after 30 days of inactivity.

@github-actions github-actions bot added the Stale label Feb 7, 2025
@github-actions github-actions bot closed this Mar 10, 2025
@github-actions github-actions bot deleted the gh/drisspg/98/head branch April 11, 2025 02:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

ciflow/inductor module: inductor Stale topic: not user facing topic category

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@drisspg @aakhundov