[inductor] fix triton bucketize mask propagation #159961

v0i0 · 2025-08-06T15:42:12Z

Stack from ghstack (oldest at bottom):

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 23c0501 Pull-Request: #159930 [ghstack-poisoned]

pytorch-bot · 2025-08-06T15:42:17Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/159961

📄 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.

⏳ 5 Pending, 2 Unrelated Failures

As of commit 08d2f87 with merge base b9003ed ():

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

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

inductor / cuda12.8-py3.10-gcc9-sm86 / test (inductor_torchbench, 1, 2, linux.g5.4xlarge.nvidia.gpu) (gh) (trunk failure)
Process completed with exit code 1.
inductor / cuda12.8-py3.10-gcc9-sm86 / test (inductor_torchbench, 2, 2, linux.g5.4xlarge.nvidia.gpu) (gh) (trunk failure)
Process completed with exit code 1.

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

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 23c0501 Pull-Request: #159930 ghstack-source-id: 5132a6b Pull Request resolved: #159961

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: Pull-Request: #159930 cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov coconutruben [ghstack-poisoned]

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: a904d7b Pull-Request: #159930 ghstack-source-id: a904d7b Pull Request resolved: #159961

v0i0 · 2025-08-06T17:07:02Z

@pytorchbot label "topic: not user facing"

davidberard98 · 2025-08-06T20:26:50Z

test/inductor/test_torchinductor.py

+        config.cpp_wrapper,
+        "cpp_wrapper samples will lead to invalid indexing",


what happens when cpp_wrapper is used - does it raise a device-side assert? (and is it during compilation or during runtime?) and do you know if that is a regression or if that is a previously-existing bug?

cpp wrapper generates examples and runs kernels using those examples. these examples have the rule - random tensor for real data, zero data for integral data (to capture the kernel cubin i think?). so if the model makes assumptions (like repeats.sum() == output_size), cpp_wrapper fails. this is why the original MR disabled the decomp for cpp_wrapper, too.

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: Pull-Request: #159930 cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov coconutruben [ghstack-poisoned]

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: Pull-Request: #159930 [ghstack-poisoned]

eellison · 2025-08-07T18:48:17Z

torch/_inductor/codegen/triton.py

@@ -2720,6 +2720,8 @@ def bucketize(
            dtype=indexing_dtype,  # type: ignore[attr-defined]
        )

+        result.mask_vars = values.mask_vars  # type: ignore[attr-defined]


Would we need to add the boundaries mask as well ?

added all CSEVariables

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: a2229df Pull-Request: #159930 ghstack-source-id: a2229df Pull Request resolved: #159961

[ghstack-poisoned]

v0i0 · 2025-08-11T23:56:34Z

@pytorchbot merge

pytorchmergebot · 2025-08-11T23:58:22Z

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 · 2025-08-11T23:58:42Z

Starting merge as part of PR stack under #158462

pytorchmergebot · 2025-08-12T02:46:45Z

Merge failed

Reason: Command git -C /home/runner/work/pytorch/pytorch cherry-pick -x 6efbbd62df8e148f48e6a0ebc50bab2240ebf6ad returned non-zero exit code 1

Auto-merging test/inductor/test_torchinductor.py
CONFLICT (content): Merge conflict in test/inductor/test_torchinductor.py
error: could not apply 6efbbd62df8... [inductor] fix triton bucketize mask propagation
hint: After resolving the conflicts, mark them with
hint: "git add/rm <pathspec>", then run
hint: "git cherry-pick --continue".
hint: You can instead skip this commit with "git cherry-pick --skip".
hint: To abort and get back to the state before "git cherry-pick",
hint: run "git cherry-pick --abort".
hint: Disable this message with "git config set advice.mergeConflict false"

Details for Dev Infra team

Raised by workflow job

[ghstack-poisoned]

v0i0 · 2025-08-12T16:25:52Z

@pytorchbot merge

pytorchmergebot · 2025-08-12T16:28:11Z

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 · 2025-08-12T16:28:17Z

Starting merge as part of PR stack under #158462

[inductor] fix triton bucketize mask propagation

eb15caf

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 23c0501 Pull-Request: #159930 [ghstack-poisoned]

pytorch-bot bot added ciflow/inductor module: inductor labels Aug 6, 2025

v0i0 requested review from eellison and davidberard98 August 6, 2025 17:05

pytorch-bot bot added the topic: not user facing topic category label Aug 6, 2025

v0i0 mentioned this pull request Aug 6, 2025

[inductor] add lowering for repeat_interleave.Tensor with output size specified (#147160) #158462

Open

davidberard98 reviewed Aug 6, 2025

View reviewed changes

v0i0 added 3 commits August 6, 2025 14:29

Update on "[inductor] fix triton bucketize mask propagation"

310e5c1

See https://hud.pytorch.org/pytorch/pytorch/commit/6b414f56a4a133a428af618d8ed1553849341497 Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: Pull-Request: #159930 [ghstack-poisoned]

eellison approved these changes Aug 7, 2025

View reviewed changes

Update

8d193cf

[ghstack-poisoned]

pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Aug 11, 2025

pytorchmergebot added the merging label Aug 11, 2025

pytorchmergebot removed the merging label Aug 12, 2025

Update

08d2f87

[ghstack-poisoned]

pytorchmergebot added the merging label Aug 12, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[inductor] fix triton bucketize mask propagation #159961

[inductor] fix triton bucketize mask propagation #159961

v0i0 commented Aug 6, 2025 •

edited by pytorch-bot bot

Loading

Uh oh!

pytorch-bot bot commented Aug 6, 2025 •

edited

Loading

Uh oh!

v0i0 commented Aug 6, 2025

Uh oh!

davidberard98 Aug 6, 2025

Uh oh!

v0i0 Aug 6, 2025

Uh oh!

eellison Aug 7, 2025

Uh oh!

v0i0 Aug 8, 2025

Uh oh!

v0i0 commented Aug 11, 2025

Uh oh!

pytorchmergebot commented Aug 11, 2025

Uh oh!

pytorchmergebot commented Aug 11, 2025

Uh oh!

pytorchmergebot commented Aug 12, 2025

Uh oh!

v0i0 commented Aug 12, 2025

Uh oh!

pytorchmergebot commented Aug 12, 2025

Uh oh!

pytorchmergebot commented Aug 12, 2025

Uh oh!

Uh oh!

		config.cpp_wrapper,
		"cpp_wrapper samples will lead to invalid indexing",

[inductor] fix triton bucketize mask propagation #159961

Are you sure you want to change the base?

[inductor] fix triton bucketize mask propagation #159961

Conversation

v0i0 commented Aug 6, 2025 • edited by pytorch-bot bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Aug 6, 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/159961

⏳ 5 Pending, 2 Unrelated Failures

Uh oh!

v0i0 commented Aug 6, 2025

Uh oh!

davidberard98 Aug 6, 2025

Choose a reason for hiding this comment

Uh oh!

v0i0 Aug 6, 2025

Choose a reason for hiding this comment

Uh oh!

eellison Aug 7, 2025

Choose a reason for hiding this comment

Uh oh!

v0i0 Aug 8, 2025

Choose a reason for hiding this comment

Uh oh!

v0i0 commented Aug 11, 2025

Uh oh!

pytorchmergebot commented Aug 11, 2025

Merge started

Uh oh!

pytorchmergebot commented Aug 11, 2025

Uh oh!

pytorchmergebot commented Aug 12, 2025

Merge failed

Uh oh!

v0i0 commented Aug 12, 2025

Uh oh!

pytorchmergebot commented Aug 12, 2025

Merge started

Uh oh!

pytorchmergebot commented Aug 12, 2025

Uh oh!

Uh oh!

v0i0 commented Aug 6, 2025 •

edited by pytorch-bot bot

Loading

pytorch-bot bot commented Aug 6, 2025 •

edited

Loading