[dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD by anijain2305 · Pull Request #162509 · pytorch/pytorch

anijain2305 · 2025-09-09T18:06:08Z

Stack from ghstack (oldest at bottom):

[dynamo][guards] Do not construct entire framelocals dict for LAMBDA_GUARD #162525
-> [dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD #162509

This is a smaller PR to reduce framelocals to dict conversion.

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames @Lucaskabela

…red LAMBDA_GUARD [ghstack-poisoned]

pytorch-bot · 2025-09-09T18:06:12Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/162509

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki

Note: Links to docs will display an error until the docs builds have been completed.

❌ 3 New Failures, 2 Unrelated Failures

As of commit 4ba3f10 with merge base 1c16c18 ():

NEW FAILURES - The following jobs have failed:

inductor / inductor-test / test (inductor_torchbench, 1, 2, linux.g5.4xlarge.nvidia.gpu) (gh)
hf_T5
inductor / unit-test / inductor-test / test (inductor_distributed, 1, 1, linux.g5.12xlarge.nvidia.gpu) (gh)
'test/distributed/test_dynamo_distributed.py::TestFakeDistributedSingleProc::test_hf_bert_ddp_aot_eager'
inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu) (gh)
'test/distributed/test_dynamo_distributed.py::TestFakeDistributedSingleProc::test_hf_bert_ddp_aot_eager'

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

inductor / inductor-cpu-test / test (inductor_torchbench_cpu_smoketest_perf, 1, 1, linux.24xl.spr-metal) (gh) (trunk failure)
Process completed with exit code 139.
inductor / inductor-test / test (inductor_huggingface, 1, 1, linux.g5.4xlarge.nvidia.gpu) (gh) (trunk failure)
Process completed with exit code 134.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

…red LAMBDA_GUARD ghstack-source-id: e3782e8 Pull Request resolved: #162509

…r not required LAMBDA_GUARD" This is a smaller PR to reduce framelocals to dict conversion. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames Lucaskabela [ghstack-poisoned]

…red LAMBDA_GUARD ghstack-source-id: b6c8bc8 Pull Request resolved: #162509

…r not required LAMBDA_GUARD" This is a smaller PR to reduce framelocals to dict conversion. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames Lucaskabela [ghstack-poisoned]

…red LAMBDA_GUARD ghstack-source-id: 10025a6 Pull Request resolved: #162509

…r not required LAMBDA_GUARD" This is a smaller PR to reduce framelocals to dict conversion. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames Lucaskabela [ghstack-poisoned]

williamwen42 · 2025-09-10T00:35:16Z

torch/csrc/dynamo/guards.cpp

+      py::object verbose_code_parts)
+      : LAMBDA_GUARD(root_guard_manager, guard_check_fn, verbose_code_parts) {}
+
+  bool check_nopybind(PyObject* value) override { // borrowed ref


How are we reaching here if the guard check always goes through

pytorch/torch/csrc/dynamo/guards.cpp

Line 6526 in e4174b1

return ((RootGuardManager*)root)->check_nopybind(f_locals);

I think this can happen in debug verbose mode, where we run guards during compilation for validation

Oh this guard is for a leaf guard. Like from_numpy, for which the the value is already accessed through a series of guard accessors.

…r not required LAMBDA_GUARD" This is a smaller PR to reduce framelocals to dict conversion. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames Lucaskabela [ghstack-poisoned]

pytorchmergebot · 2025-09-10T18:46:06Z

Starting merge as part of PR stack under #162525

…GUARD (#162525) Pull Request resolved: #162525 Approved by: https://github.com/williamwen42 ghstack dependencies: #162509

…red LAMBDA_GUARD (pytorch#162509) This is a smaller PR to reduce framelocals to dict conversion. Pull Request resolved: pytorch#162509 Approved by: https://github.com/williamwen42

…GUARD (pytorch#162525) Pull Request resolved: pytorch#162525 Approved by: https://github.com/williamwen42 ghstack dependencies: pytorch#162509

anijain2305 · 2025-09-19T23:28:08Z

@pytorchbot revert -m "Reverting because fburl.com/workplace/qiitdlp6 is prompting us to think about an alternate implementation" -c nosignal

pytorchmergebot · 2025-09-19T23:30:16Z

@pytorchbot successfully started a revert job. Check the current status here.
Questions? Feedback? Please reach out to the PyTorch DevX Team

pytorchmergebot · 2025-09-19T23:30:28Z

Reverting PR 162509 failed

Reason: Command git -C /home/runner/work/pytorch/pytorch revert --no-edit a67e798cb79a9633d89234f3316e4f165b9c8d5a returned non-zero exit code 1

Auto-merging torch/_C/_dynamo/guards.pyi
Auto-merging torch/_dynamo/guards.py
CONFLICT (content): Merge conflict in torch/_dynamo/guards.py
Auto-merging torch/csrc/dynamo/guards.cpp
error: could not revert a67e798cb79... [dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD (#162509)
hint: After resolving the conflicts, mark them with
hint: "git add/rm <pathspec>", then run
hint: "git revert --continue".
hint: You can instead skip this commit with "git revert --skip".
hint: To abort and get back to the state before "git revert",
hint: run "git revert --abort".
hint: Disable this message with "git config set advice.mergeConflict false"

Details for Dev Infra team

Raised by workflow job

…red LAMBDA_GUARD (pytorch#162509) This is a smaller PR to reduce framelocals to dict conversion. Pull Request resolved: pytorch#162509 Approved by: https://github.com/williamwen42

…GUARD (pytorch#162525) Pull Request resolved: pytorch#162525 Approved by: https://github.com/williamwen42 ghstack dependencies: pytorch#162509

…red LAMBDA_GUARD (pytorch#162509) This is a smaller PR to reduce framelocals to dict conversion. Pull Request resolved: pytorch#162509 Approved by: https://github.com/williamwen42

…GUARD (pytorch#162525) Pull Request resolved: pytorch#162525 Approved by: https://github.com/williamwen42 ghstack dependencies: pytorch#162509

…red LAMBDA_GUARD (pytorch#162509) This is a smaller PR to reduce framelocals to dict conversion. Pull Request resolved: pytorch#162509 Approved by: https://github.com/williamwen42

…GUARD (pytorch#162525) Pull Request resolved: pytorch#162525 Approved by: https://github.com/williamwen42 ghstack dependencies: pytorch#162509

[dynamo][guards] Prevent framelocals to dict conversion for not requi…

8b48356

…red LAMBDA_GUARD [ghstack-poisoned]

pytorch-bot bot added ciflow/inductor module: dynamo labels Sep 9, 2025

anijain2305 added a commit that referenced this pull request Sep 9, 2025

[dynamo][guards] Prevent framelocals to dict conversion for not requi…

e093270

…red LAMBDA_GUARD ghstack-source-id: e3782e8 Pull Request resolved: #162509

anijain2305 added ciflow/trunk Trigger trunk jobs on your pull request topic: not user facing topic category labels Sep 9, 2025

anijain2305 added a commit that referenced this pull request Sep 9, 2025

[dynamo][guards] Prevent framelocals to dict conversion for not requi…

d214455

…red LAMBDA_GUARD ghstack-source-id: b6c8bc8 Pull Request resolved: #162509

anijain2305 mentioned this pull request Sep 9, 2025

[dynamo][refactor] Move get_framelocals_idx to a helper #162519

Closed

anijain2305 added a commit that referenced this pull request Sep 9, 2025

[dynamo][guards] Prevent framelocals to dict conversion for not requi…

2d77dc3

…red LAMBDA_GUARD ghstack-source-id: 10025a6 Pull Request resolved: #162509

anijain2305 mentioned this pull request Sep 9, 2025

[dynamo][guards] Do not construct entire framelocals dict for LAMBDA_GUARD #162525

Closed

anijain2305 requested a review from williamwen42 September 9, 2025 22:55

williamwen42 reviewed Sep 10, 2025

View reviewed changes

anijain2305 requested a review from williamwen42 September 10, 2025 05:55

anijain2305 assigned jansel Sep 10, 2025

anijain2305 requested a review from jansel September 10, 2025 16:43

anijain2305 unassigned jansel Sep 10, 2025

williamwen42 approved these changes Sep 10, 2025

View reviewed changes

pytorchmergebot closed this in a67e798 Sep 10, 2025

pytorchmergebot added the Merged label Sep 10, 2025

pytorchmergebot pushed a commit that referenced this pull request Sep 10, 2025

[dynamo][guards] Do not construct entire framelocals dict for LAMBDA_…

5f630d2

…GUARD (#162525) Pull Request resolved: #162525 Approved by: https://github.com/williamwen42 ghstack dependencies: #162509

anijain2305 mentioned this pull request Sep 10, 2025

[profile trace] Fill gaps in compile time profile traces #162627

Closed

github-actions bot deleted the gh/anijain2305/867/head branch October 20, 2025 02:17

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD#162509

[dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD#162509
anijain2305 wants to merge 5 commits intogh/anijain2305/867/basefrom
gh/anijain2305/867/head

anijain2305 commented Sep 9, 2025 •

edited

Loading

Uh oh!

pytorch-bot bot commented Sep 9, 2025 •

edited

Loading

Uh oh!

williamwen42 Sep 10, 2025

Uh oh!

anijain2305 Sep 10, 2025

Uh oh!

anijain2305 Sep 10, 2025

Uh oh!

pytorchmergebot commented Sep 10, 2025

Uh oh!

anijain2305 commented Sep 19, 2025

Uh oh!

pytorchmergebot commented Sep 19, 2025

Uh oh!

pytorchmergebot commented Sep 19, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

anijain2305 commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/162509

❌ 3 New Failures, 2 Unrelated Failures

Uh oh!

williamwen42 Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

anijain2305 Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

anijain2305 Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

pytorchmergebot commented Sep 10, 2025

Uh oh!

anijain2305 commented Sep 19, 2025

Uh oh!

pytorchmergebot commented Sep 19, 2025

Uh oh!

pytorchmergebot commented Sep 19, 2025

Reverting PR 162509 failed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

anijain2305 commented Sep 9, 2025 •

edited

Loading

pytorch-bot bot commented Sep 9, 2025 •

edited

Loading