[inductor] parallel compile: Create new pipes for subproc communication #133590
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
…on (#131194) Summary: Rather then using stdin/stdout for IPC, we can create new pipes and pass the descriptors to the subproc via the cmd line. #131070 reports an issue where the combination of deepspeed and onnxruntime-training causes _something_ in the subproc to write to stdout and corrupt the IPC. The current implementation was already brittle; we can just create new pipes specifically for the IPC. Test Plan: I was able to repro the MemoryError in #131070 by installing deepspeed and onnxruntime-training. Verified this PR fixes. Differential Revision: [D59968362](https://our.internmc.facebook.com/intern/diff/D59968362) Pull Request resolved: #131194 Approved by: https://github.com/malfet, https://github.com/eellison, https://github.com/atalman (cherry picked from commit 3c43fe0)
This was referenced Aug 15, 2024
pytorchmergebot
pushed a commit
that referenced
this pull request
Aug 20, 2024
During cherry-picking we want to use default setting and fail if there is merge conflict Here an example of invalid conflict resolution: #131194 and cherry-pick #133590 Pull Request resolved: #134047 Approved by: https://github.com/kit1980
malfet
pushed a commit
to aditew01/pytorch
that referenced
this pull request
Sep 13, 2024
During cherry-picking we want to use default setting and fail if there is merge conflict Here an example of invalid conflict resolution: pytorch#131194 and cherry-pick pytorch#133590 Pull Request resolved: pytorch#134047 Approved by: https://github.com/kit1980
|
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.
Stack from ghstack (oldest at bottom):
Summary: Rather then using stdin/stdout for IPC, we can create new pipes and pass the descriptors to the subproc via the cmd line. #131070 reports an issue where the combination of deepspeed and onnxruntime-training causes something in the subproc to write to stdout and corrupt the IPC. The current implementation was already brittle; we can just create new pipes specifically for the IPC.
Test Plan: I was able to repro the MemoryError in #131070 by installing deepspeed and onnxruntime-training. Verified this PR fixes.
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang @aakhundov
Differential Revision: D59968362