Improved unbacked SymInt input support in Inductor #124739
Closed
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
This was referenced Apr 23, 2024
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/124739
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 7205be7 with merge base a22847a ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Apr 23, 2024
This was referenced Apr 23, 2024
Closed
github-actions
bot
requested review from
albanD,
antoniojkim,
bdhirsh,
miladm and
SherlockNoMad
Closed
ezyang
requested review from
jansel,
lezcano,
avikchaudhuri,
IvanKobzarev and
ColinPeppler
Closed
This was referenced Apr 23, 2024
This was referenced Apr 24, 2024
jansel
approved these changes
Apr 25, 2024
|
@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 |
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 25, 2024
#124782) Previously, unbacked SymInts would gradually get larger and larger as we kept rebinding them. Now, we do the replacement to preserve the old symbol. Actually doing this is a bit tricky. Here’s the order things happen when retracing data dependent: 1. Run fake tensor prop: allocate new unbacked SymInt 2. Run proxy tensor mode, calculate bindings and associate them with FX node 3. Run PropagateUnbackedSymInts, rename unbacked bindings to their old ones so they are consistent So the problem is when we calculate bindings in step (2), we don't know what the original names are yet, we only find out later at (3). But by the time (3) runs, we've already stuffed some new bindings in meta["unbacked_bindings"] and we don't know how to update them! To fix this, I introduce resolve_unbacked_bindings which post facto applies any of the renamings we discovered in (3). Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: #124782 Approved by: https://github.com/lezcano ghstack dependencies: #124310, #124314, #124316, #124394, #124739
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 25, 2024
…ass (#123735) This ensures that first argument to record_shapeenv_event is a ShapeEnv so we can appropriately short circuit when recording is not in progress. Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: #123735 Approved by: https://github.com/ysiraichi, https://github.com/zou3519, https://github.com/albanD ghstack dependencies: #124310, #124314, #124316, #124394, #124739, #124782, #124785
pytorch-bot bot
pushed a commit
that referenced
this pull request
May 3, 2024
This is a subset of changes extracted from #124683 This PR contains modifications to make Inductor work with unbacked symbol inputs, which can occur when a data-dependent sized tensor is saved for backwards. The problems to be fixed: * When binding initial symbols, we unconditionally bind unbacked symbols (instead of computing if they are needed, which only looks at backed symbols) * Benchmark generation code doesn't work with unbacked symints as we have no hints to actually feed in real values. So I pick a random number and you are expected to fix it if it doesn't work * Need to make sure we don't install dependencies on unbacked SymInt inputs, that puts us down the "promptly deallocate the input" path, but that's pointless for unbacked SymInt Fixes #124652 Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: #124739 Approved by: https://github.com/jansel ghstack dependencies: #124310, #124314, #124316, #124394
petrex
pushed a commit
to petrex/pytorch
that referenced
this pull request
May 3, 2024
pytorch#124782) Previously, unbacked SymInts would gradually get larger and larger as we kept rebinding them. Now, we do the replacement to preserve the old symbol. Actually doing this is a bit tricky. Here’s the order things happen when retracing data dependent: 1. Run fake tensor prop: allocate new unbacked SymInt 2. Run proxy tensor mode, calculate bindings and associate them with FX node 3. Run PropagateUnbackedSymInts, rename unbacked bindings to their old ones so they are consistent So the problem is when we calculate bindings in step (2), we don't know what the original names are yet, we only find out later at (3). But by the time (3) runs, we've already stuffed some new bindings in meta["unbacked_bindings"] and we don't know how to update them! To fix this, I introduce resolve_unbacked_bindings which post facto applies any of the renamings we discovered in (3). Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: pytorch#124782 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#124310, pytorch#124314, pytorch#124316, pytorch#124394, pytorch#124739
petrex
pushed a commit
to petrex/pytorch
that referenced
this pull request
May 3, 2024
…ytorch#124785) This ensures we have correct SymInts when we allocate tangents. Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: pytorch#124785 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#124310, pytorch#124314, pytorch#124316, pytorch#124394, pytorch#124739, pytorch#124782
pytorch-bot bot
pushed a commit
that referenced
this pull request
May 3, 2024
…ass (#123735) This ensures that first argument to record_shapeenv_event is a ShapeEnv so we can appropriately short circuit when recording is not in progress. Signed-off-by: Edward Z. Yang <ezyang@meta.com> Pull Request resolved: #123735 Approved by: https://github.com/ysiraichi, https://github.com/zou3519, https://github.com/albanD ghstack dependencies: #124310, #124314, #124316, #124394, #124739, #124782, #124785
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):
This is a subset of changes extracted from #124683
This PR contains modifications to make Inductor work with unbacked symbol inputs, which can occur when a data-dependent sized tensor is saved for backwards. The problems to be fixed:
Fixes #124652
Signed-off-by: Edward Z. Yang ezyang@meta.com
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang