autograd codegen: bump VC properly for mutable ops with no returns #133044
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
[ghstack-poisoned]
github-actions
bot
requested review from
SherlockNoMad,
antoniojkim,
ezyang and
miladm
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. [ghstack-poisoned]
bdhirsh
commented
Aug 9, 2024
| at::TensorList tensors = {first, first}; | ||
| at::TensorList undefined_tensors = {first, second}; | ||
| at::TensorList steps = {step, step}; | ||
| return at::_fused_adamw_(tensors, tensors, tensors, tensors, undefined_tensors, |
Contributor
Author
There was a problem hiding this comment.
so calling _fused_adamw_ here with undefined inputs feels like it was just wrong (we now error because you are passing in a mutable, undefined tensor, which autograd is not allowed to bump the VC on).
To make the test a bit more realistic I updated it to use aten::index(). which accepts undefined tensor indices (which are not mutated)
Contributor
Author
|
@pytorchbot merge |
Collaborator
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 |
Collaborator
Merge failedReason: 1 mandatory check(s) failed. The first few are: Dig deeper by viewing the failures on hud |
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. [ghstack-poisoned]
Contributor
|
@pytorchbot merge |
Collaborator
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 |
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
… returns" Fixes #132014 In particular, this PR should: (1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like `split_with_sizes_copy.out`) (2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names. cc XilunWu H-Huang awgu kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames rec [ghstack-poisoned]
Contributor
|
Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #132014
In particular, this PR should:
(1) start generating and registering kernels to the ADInplaceOrView dispatch key, for mutable ops that have no returns (like
split_with_sizes_copy.out)(2) prevously, the codegen would loop over mutable return names to decide which tensors to codegen VC bumps for. Now, we just loop over mutable input argument names.
Stack from ghstack (oldest at bottom):
cc @XilunWu @H-Huang @awgu @kwen2501 @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @c-p-i-o @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames @rec