DeepCompile: Specify tensor aliasing in C++ op schema #7597
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
PyTorch C++ op schema [1] allows specifying tensor storage aliasing by annotating `(a)` after input/output types. Torch inductor takes this information to determine where to insert explicit `del` statements for tensors that are no longer needed. If what an op schema specifies disagrees with the op implementation, inductor-generated code is likely to release tensors earlier than expected and leads to wrong results. `wait_allgather` and `release_param` return the first argument unchanged and that aliasing should be annotated in the schema. Also remove the code related to `clone_custom_op_output` as it is solely a workaround of the aforementioned issue. [1] https://github.com/pytorch/pytorch/blob/main/aten/src/ATen/native/README.md Fixes: deepspeedai#7596 Signed-off-by: Junjie Mao <[email protected]>
This was referenced Sep 26, 2025
tohtana
approved these changes
Sep 28, 2025
Collaborator
tohtana
left a comment
tohtana
left a comment
There was a problem hiding this comment.
Wow, amazing finding! I really appreciate your deep investigation.
Thank you for the great fix!
mauryaavinash95
pushed a commit
to DataStates/DeepSpeed
that referenced
this pull request
Oct 4, 2025
PyTorch C++ op schema [1] allows specifying tensor storage aliasing by annotating `(a)` after input/output types. Torch inductor takes this information to determine where to insert explicit `del` statements for tensors that are no longer needed. If what an op schema specifies disagrees with the op implementation, inductor-generated code is likely to release tensors earlier than expected and leads to wrong results. `wait_allgather` and `release_param` return the first argument unchanged and that aliasing should be annotated in the schema. Also remove the code related to `clone_custom_op_output` as it is solely a workaround of the aforementioned issue. [1] https://github.com/pytorch/pytorch/blob/main/aten/src/ATen/native/README.md Fixes: deepspeedai#7596 Signed-off-by: Junjie Mao <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
PyTorch C++ op schema [1] allows specifying tensor storage aliasing by annotating
(a)after input/output types. Torch inductor takes this information to determine where to insert explicitdelstatements for tensors that are no longer needed.If what an op schema specifies disagrees with the op implementation, inductor-generated code is likely to release tensors earlier than expected and leads to wrong results.
wait_allgatherandrelease_paramreturn the first argument unchanged and that aliasing should be annotated in the schema.Also remove the code related to
clone_custom_op_outputas it is solely a workaround of the aforementioned issue.[1] https://github.com/pytorch/pytorch/blob/main/aten/src/ATen/native/README.md
Fixes: #7596