[Inductor] Allow matmul to have flexiable layout when we are not autotuning #110726
Conversation
…tuning Fixes #102804 [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/110726
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (2 Unrelated Failures)As of commit 918e9d0 with merge base e8f1f4e ( UNSTABLE - The following jobs failed but were likely due to flakiness present on trunk and has been marked as unstable:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
| "Force channels last inputs for %d conv for the current graph with id %d", | ||
| self.num_channels_last_conv, | ||
| self.graph_id, | ||
| self.graph_id if self.graph_id is not None else -1, |
There was a problem hiding this comment.
None was making %d crash under TORCH_LOGS="+inductor"
…re not autotuning" Fixes #102804 cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy yf225 chenyang78 kadeng muchulee8 aakhundov ColinPeppler [ghstack-poisoned]
|
|
||
| if ( | ||
| len(choices) == 1 | ||
| and use_aten_gemm_kernels() |
There was a problem hiding this comment.
I guess it's probably a bit cleaner to just remove this condition?
There was a problem hiding this comment.
Do you mean remove the use_aten_gemm_kernels? We would also remove the restriction above then, right? If it safe to do so, i can remove both
There was a problem hiding this comment.
The idea is that if you only have one kernel choice, then the layout choice won't have any affect on which kernel gets chosen. So it's less likely to matter as much.
|
@pytorchbot merge |
Merge failedReason: This PR needs a If not, please add the To add a label, you can comment to pytorchbot, for example For more information, see Details for Dev Infra teamRaised by workflow job |
|
@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 |
4 participants
Stack from ghstack (oldest at bottom):
Fixes #102804
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @Xia-Weiwen @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @aakhundov @ColinPeppler