use statically known true instead of guard size oblivious in bmm and mm inductor decompositions . #148893
Conversation
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/148893
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit b68c0ef with merge base 0f8613b ( 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. |
…livous" [ghstack-poisoned]
…livous" [ghstack-poisoned]
…in bmm and mm decompositions . " this was discussed with eellison and he recommended using statically_known_true here, the intuition is. We already have 0/1 specializations in place, if we reach those checks with dynamic shapes that are not already specialized then we do not want them to specialize them, "a recompilation here is not justified". Those are all non-semantic changing optimizations. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov [ghstack-poisoned]
ghstack-source-id: 992156c Pull Request resolved: pytorch/pytorch#148893
There was a problem hiding this comment.
I'm a little worried this will regress performance in gpt-fast, but, I agree that today this is not the best experience.
Getting these firing will be a nice end result of @bobrenjc93's work
|
We can always go back to guard_or_false if this end up being a regression @eellison |
|
@pytorchbot merge |
Merge startedYour 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 |
|
The merge job was canceled or timed out. This most often happen if two merge requests were issued for the same PR, or if merge job was waiting for more than 6 hours for tests to finish. In later case, please do not hesitate to reissue the merge command |
|
@pytorchbot merge -f "GH is broken" |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Stack from ghstack (oldest at bottom):
this was discussed with @eellison and he recommended using statically_known_true here, the intuition is. We already have 0/1 specializations in place, if we reach those checks with dynamic shapes that are not already specialized
then we do not want them to specialize them, "a recompilation here is not justified".
Those are all non-semantic changing optimizations.
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov