Remove guard_size_oblivious from vector_norm decomposition. #148809

laithsakka · 2025-03-08T04:04:43Z

Stack from ghstack (oldest at bottom):

-> Remove guard_size_oblivious from vector_norm decomposition. #148809

This PR remove the usage of guard_size_oblivious in vector_norm by inlining it in the runtime check,
this prevent any data dependent error from ever appearing here at the locations where guard_size_oblivious
used to exist. Before this PR it used to break potentially. This is NOT BC breaking or changing of semantics from eager.

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

[ghstack-poisoned]

pytorch-bot · 2025-03-08T04:04:46Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/148809

📄 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 or our office hours

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

❌ 6 New Failures, 2 Unrelated Failures

As of commit e13a5f0 with merge base 6841451 ():

NEW FAILURES - The following jobs have failed:

linux-binary-libtorch-release / libtorch-cpu-shared-with-deps-release-build / build (gh)
Process completed with exit code 1.
linux-binary-manywheel / manywheel-py3_9-cuda11_8-build / build (gh)
Process completed with exit code 1.
linux-binary-manywheel / manywheel-py3_9-cuda12_6-build / build (gh)
Process completed with exit code 1.
linux-binary-manywheel / manywheel-py3_9-cuda12_8-build / build (gh)
Process completed with exit code 1.
pull / cuda12.4-py3.10-gcc9-sm75 / test (pr_time_benchmarks, 1, 1, linux.g4dn.metal.nvidia.gpu) (gh)
REGRESSION: benchmark ('aotdispatcher_partitioner_cpu2', 'compile_time_instruction_count') failed, actual result 1751440693 is 2.07% higher than expected 1716000000 ±+1.50% if this is an expected regression, please update the expected results.
pull / linux-jammy-py3.9-gcc11 / test (backwards_compat, 1, 1, lf.ephemeral.linux.2xlarge) (gh)
test_modules_can_be_imported

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

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

pull / linux-focal-py3.13-clang10 / test (dynamo_wrapped, 3, 3, lf.ephemeral.linux.2xlarge) (gh) (trunk failure)
test_nestedtensor.py::TestNestedTensorSubclassCPU::test_composite_op_in_inference_mode_cpu_float32
pull / linux-focal-py3.9-clang10 / test (dynamo_wrapped, 3, 3, lf.ephemeral.linux.2xlarge) (gh) (trunk failure)
test_nestedtensor.py::TestNestedTensorSubclassCPU::test_composite_op_in_inference_mode_cpu_float32

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

[ghstack-poisoned]

ghstack-source-id: 95d4cf1 Pull Request resolved: #148809

[ghstack-poisoned]

ghstack-source-id: 8ca3454 Pull Request resolved: pytorch/pytorch#148809

[ghstack-poisoned]

bobrenjc93 · 2025-04-04T01:21:56Z

torch/fx/experimental/symbolic_shapes.py

+# When the whole expression is passed to evaluation in that case, we do not throw a
+# data dependent error or guard because we can statically know the result is True
+# before unpacking the symbols.
+sym_or = operator.or_


given that @pianpwk is introducing sym_or in #150456 (comment), does it make sense to rebase on top of his PR?

he can rebase on top of mine and remove this maybe
i put this PR long ago and want to land it . @PiaN do you mind ?

also maybe add this comment on your PR on the new function once you rebase @pianpwk

laithsakka · 2025-04-04T01:25:17Z

@pytorchbot merge

This PR remove the usage of guard_size_oblivious in vector_norm by inlining it in the runtime check, this prevent any data dependent error from ever appearing here at the locations where guard_size_oblivious used to exist. Before this PR it used to break potentially. This is NOT BC breaking or changing of semantics from eager. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx chenyang78 kadeng chauhang amjames [ghstack-poisoned]

ghstack-source-id: b8c2472 Pull Request resolved: #148809

pytorchmergebot · 2025-04-04T01:26:58Z

Merge started

Your 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

Advanced Debugging

Check the merge workflow status
here

laithsakka · 2025-04-04T01:27:05Z

@pytorchbot merge

pytorchmergebot · 2025-04-04T01:27:23Z

The merge job was canceled or timed out. This most often happen if two merge requests were issued for the same PR, or if merge job was waiting for more than 6 hours for tests to finish. In later case, please do not hesitate to reissue the merge command
For more information see pytorch-bot wiki.

pytorchmergebot · 2025-04-04T01:29:25Z

Merge started

Your 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

Advanced Debugging

Check the merge workflow status
here

pytorchmergebot · 2025-04-04T01:45:19Z

Merge failed

Reason: 3 jobs have failed, first few of them are: linux-binary-manywheel / manywheel-py3_9-cuda12_6-build / build, linux-binary-manywheel / manywheel-py3_9-cuda11_8-build / build, linux-binary-libtorch-release / libtorch-cpu-shared-with-deps-release-build / build

Details for Dev Infra team

Raised by workflow job

laithsakka · 2025-04-04T15:54:26Z

@pytorchbot merge

pytorchmergebot · 2025-04-04T15:56:26Z

Merge started

Your 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

Advanced Debugging

Check the merge workflow status
here

pytorchmergebot · 2025-04-04T15:56:35Z

Merge failed

Reason: 1 mandatory check(s) failed. The first few are:

pull / linux-jammy-py3.9-gcc11 / test (backwards_compat, 1, 1, lf.ephemeral.linux.2xlarge)

Dig deeper by viewing the failures on hud

Details for Dev Infra team

Raised by workflow job

Failing merge rule: Core Maintainers

laithsakka · 2025-04-10T16:11:27Z

@pytorchbot merge -i

pytorchmergebot · 2025-04-10T16:13:20Z

Merge started

Your change will be merged while ignoring the following 8 checks: pull / cuda12.4-py3.10-gcc9-sm75 / test (pr_time_benchmarks, 1, 1, linux.g4dn.metal.nvidia.gpu), pull / linux-focal-py3.13-clang10 / test (dynamo_wrapped, 3, 3, lf.ephemeral.linux.2xlarge), pull / linux-jammy-py3.9-gcc11 / test (backwards_compat, 1, 1, lf.ephemeral.linux.2xlarge), pull / linux-focal-py3.9-clang10 / test (dynamo_wrapped, 3, 3, lf.ephemeral.linux.2xlarge), linux-binary-manywheel / manywheel-py3_9-cuda12_6-build / build, linux-binary-manywheel / manywheel-py3_9-cuda11_8-build / build, linux-binary-manywheel / manywheel-py3_9-cuda12_8-build / build, linux-binary-libtorch-release / libtorch-cpu-shared-with-deps-release-build / build

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

…#148809) This PR remove the usage of guard_size_oblivious in vector_norm by inlining it in the runtime check, this prevent any data dependent error from ever appearing here at the locations where guard_size_oblivious used to exist. Before this PR it used to break potentially. This is NOT BC breaking or changing of semantics from eager. Pull Request resolved: pytorch#148809 Approved by: https://github.com/bobrenjc93

ghstack-source-id: f1df685 Pull Request resolved: pytorch/pytorch#148809

remove gaurd_size_oblivious from vector_norm

9d9180a

[ghstack-poisoned]

This was referenced Mar 8, 2025

Introduce guard_or_true, guard_or_false #148430

Closed

avoid guard_size_oblivious in vector_norm #148810

Closed

Update on "remove gaurd_size_oblivious from vector_norm"

082ecce

[ghstack-poisoned]

laithsakka changed the title ~~remove gaurd_size_oblivious from vector_norm~~ [DRAFT] guard_size_oblivious from vector_norm Mar 8, 2025

laithsakka changed the title ~~[DRAFT] guard_size_oblivious from vector_norm~~ [DRAFT] remove guard_size_oblivious from vector_norm Mar 8, 2025

Update on "[DRAFT] remove guard_size_oblivious from vector_norm"

4110efb

[ghstack-poisoned]

laithsakka added a commit that referenced this pull request Mar 8, 2025

remove gaurd_size_oblivious from vector_norm

a39aaa3

ghstack-source-id: 95d4cf1 Pull Request resolved: #148809

laithsakka mentioned this pull request Mar 16, 2025

[Pushed as separate PR on different stack] cache kernel codegen and loading #149266

Closed

Update on "[DRAFT] remove guard_size_oblivious from vector_norm"

e453d32

[ghstack-poisoned]

jurgen-paul pushed a commit to jurgen-paul/pytorch.git.file that referenced this pull request Mar 19, 2025

remove gaurd_size_oblivious from vector_norm

36e2f20

ghstack-source-id: 8ca3454 Pull Request resolved: pytorch/pytorch#148809

Update on "[DRAFT] remove guard_size_oblivious from vector_norm"

51585ed

[ghstack-poisoned]

laithsakka changed the title ~~[DRAFT] remove guard_size_oblivious from vector_norm~~ remove guard_size_oblivious from vector_norm Mar 25, 2025

laithsakka changed the title ~~remove guard_size_oblivious from vector_norm~~ Remove guard_size_oblivious from vector_norm Mar 25, 2025

laithsakka requested review from aorenste, avikchaudhuri, bdhirsh, bobrenjc93, ezyang and pianpwk and removed request for ezyang March 25, 2025 07:59

Update on " Remove guard_size_oblivious from vector_norm"

fb181c1

[ghstack-poisoned]

bobrenjc93 reviewed Apr 4, 2025

View reviewed changes

bobrenjc93 approved these changes Apr 4, 2025

View reviewed changes

pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Apr 4, 2025

laithsakka added a commit that referenced this pull request Apr 4, 2025

remove gaurd_size_oblivious from vector_norm

f00a692

ghstack-source-id: b8c2472 Pull Request resolved: #148809

pytorchmergebot added the merging label Apr 4, 2025

pytorchmergebot removed the merging label Apr 4, 2025

pytorchmergebot added the merging label Apr 4, 2025

pytorchmergebot removed the merging label Apr 4, 2025

pytorchmergebot added the merging label Apr 10, 2025

pytorchmergebot closed this in 5471e80 Apr 10, 2025

pytorchmergebot added Merged and removed merging labels Apr 10, 2025

Divigroup-RAP pushed a commit to Divigroup-RAP/PYTORCH that referenced this pull request Apr 22, 2025

remove gaurd_size_oblivious from vector_norm

f4d6dd6

ghstack-source-id: f1df685 Pull Request resolved: pytorch/pytorch#148809

github-actions bot deleted the gh/laithsakka/115/head branch May 15, 2025 02:16

AlonSardas mentioned this pull request May 21, 2025

torch.compile does not support lambda error message arguments for torch._check #149997

Open

Remove guard_size_oblivious from vector_norm decomposition. #148809

Remove guard_size_oblivious from vector_norm decomposition. #148809

Uh oh!

Conversation

laithsakka commented Mar 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Mar 8, 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/148809

❌ 6 New Failures, 2 Unrelated Failures

Uh oh!

bobrenjc93 Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

laithsakka Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

laithsakka Apr 4, 2025

Choose a reason for hiding this comment

Uh oh!

laithsakka commented Apr 4, 2025

Uh oh!

pytorchmergebot commented Apr 4, 2025

Merge started

Uh oh!

laithsakka commented Apr 4, 2025

Uh oh!

pytorchmergebot commented Apr 4, 2025

Uh oh!

pytorchmergebot commented Apr 4, 2025

Merge started

Uh oh!

pytorchmergebot commented Apr 4, 2025

Merge failed

Uh oh!

laithsakka commented Apr 4, 2025

Uh oh!

pytorchmergebot commented Apr 4, 2025

Merge started

Uh oh!

pytorchmergebot commented Apr 4, 2025

Merge failed

Uh oh!

laithsakka commented Apr 10, 2025

Uh oh!

pytorchmergebot commented Apr 10, 2025

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

laithsakka commented Mar 8, 2025 •

edited

Loading

pytorch-bot bot commented Mar 8, 2025 •

edited

Loading

bobrenjc93 Apr 4, 2025 •

edited

Loading

laithsakka Apr 4, 2025 •

edited

Loading