Add torch._C.Tag.needs_contiguous_strides #152859
Conversation
this forces inductor to force the inputs to be contiguous. Test Plan: - new test [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/152859
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 f3a59f0 with merge base fe36d7d ( UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
this forces inductor to force the inputs to be contiguous. Test Plan: - new test cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov [ghstack-poisoned]
torch/_inductor/lowering.py
Outdated
| @@ -164,6 +164,8 @@ def maybe_layout_constraints(fn: Callable[..., Any]) -> Optional[Callable[..., A | |||
| def tag_to_layout_constraint(tag): | |||
| if tag == torch._C.Tag.needs_exact_strides: | |||
| return constrain_to_fake_tensors | |||
| if tag == torch._C.Tag.needs_contiguous_strides: | |||
| return require_contiguous | |||
There was a problem hiding this comment.
Can we construct the strides manually ? I had a pr to fix this but it was blocked by mkldnn tests. @CaoE has a pr to update #148424. cc @CaoE @leslie-fang-intel we really need to land that.
There was a problem hiding this comment.
What do you mean by "Can we construct the strides manually ?"?
this forces inductor to force the inputs to be contiguous. Test Plan: - new test cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov [ghstack-poisoned]
|
@pytorchbot merge -i |
Merge startedYour change will be merged while ignoring the following 1 checks: pull / unstable-linux-focal-cuda12.6-py3.10-gcc11-sm89-xfail / build Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Stack from ghstack (oldest at bottom):
this forces inductor to force the inputs to be contiguous.
Test Plan:
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov