Support negative index slicing with backed symints by ColinPeppler · Pull Request #177308 · pytorch/pytorch

ColinPeppler · 2026-03-12T20:28:27Z

Stack from ghstack (oldest at bottom):

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

[ghstack-poisoned]

pytorch-bot · 2026-03-12T20:28:31Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/177308

📄 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.

✅ No Failures

As of commit 907a318 with merge base a345892 ():
💚 Looks good so far! There are no failures yet. 💚

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

pytorch-bot · 2026-03-12T20:28:34Z

This PR needs a `release notes:` label

If your changes are user facing and intended to be a part of release notes, please use a label starting with release notes:.

If not, please add the topic: not user facing label.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "topic: not user facing"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

ghstack-source-id: b6111aa Pull Request resolved: #177308

laithsakka · 2026-03-12T22:30:43Z

test/inductor/test_aot_inductor.py

+                shifts = torch.arange(0, 64, 8, device=x.device, dtype=torch.int64)
+                return (expanded >> shifts) & 255
+
+        torch.cuda.caching_allocator_enable(False)


ditto same as previous PR pls address before landing do try catch finally maybe?

laithsakka · 2026-03-12T22:43:01Z

torch/_inductor/ir.py

        if any(free_unbacked_symbols(x) for x in (start, end, dim_size)):
            min_func = sympy.Min
            max_func = sympy.Max
+        elif any(


mm this is only needed when backed_size_oblivious is on .

torch/_subclasses/fake_impls.py

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler · 2026-03-19T03:04:44Z

@pytorchbot merge

pytorchmergebot · 2026-03-19T03:06:47Z

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

This reverts commit 9a7ae22. Reverted #177308 on behalf of https://github.com/yangw-dev due to sorry the pr breaks internal test RiskExtrapolationModuleStabilityTest, please fix it and reland D97174518 ([comment](#175819 (comment)))

pytorchmergebot · 2026-03-19T14:58:56Z

@ColinPeppler your PR has been reverted as part of the stack under #175819.

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka ghstack dependencies: pytorch#175819

…77308)" This reverts commit 9a7ae22. Reverted pytorch#177308 on behalf of https://github.com/yangw-dev due to sorry the pr breaks internal test RiskExtrapolationModuleStabilityTest, please fix it and reland D97174518 ([comment](pytorch#175819 (comment)))

pytorchmergebot · 2026-03-30T23:23:51Z

Starting merge as part of PR stack under #177418

pytorchmergebot · 2026-03-31T01:25:58Z

Starting merge as part of PR stack under #177418

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka ghstack dependencies: pytorch#175819

…77308)" This reverts commit 9a7ae22. Reverted pytorch#177308 on behalf of https://github.com/yangw-dev due to sorry the pr breaks internal test RiskExtrapolationModuleStabilityTest, please fix it and reland D97174518 ([comment](pytorch#175819 (comment)))

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler · 2026-04-02T19:11:38Z

@pytorchbot rebase -b main

pytorchmergebot · 2026-04-02T19:13:21Z

@pytorchbot started a rebase job onto refs/remotes/origin/main. Check the current status here

[ghstack-poisoned]

pytorchmergebot · 2026-04-02T19:13:37Z

Successfully rebased gh/ColinPeppler/4/orig onto refs/remotes/origin/main, please pull locally before adding more changes (for example, via ghstack checkout https://github.com/pytorch/pytorch/pull/177308)

ghstack-source-id: 8f31f7c Pull Request resolved: #177308

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

pytorchmergebot · 2026-04-03T23:58:08Z

Starting merge as part of PR stack under #177418

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 130382e Pull Request resolved: #177308

pytorchmergebot · 2026-04-07T16:59:36Z

Starting merge as part of PR stack under #177418

#177418) ### Why - An IMA debugging aid to specifically disable CCA on a targeted block of code. - Another option is `PYTORCH_NO_CUDA_MEMORY_CACHING=1` but that is set globally. Usually I'd do this. ``` torch.cuda.caching_allocator_enable(False) try: ... finally: # make sure to clean up even on exception torch.cuda.caching_allocator_enable(True) ``` ### What Add a utility that - Disables CUDA caching allocator (CCA) when entering the block. - Restores the CCA state when exiting the block (even on exceptions). ``` with torch.cuda.caching_allocator_disabled(): ... ``` Pull Request resolved: #177418 Approved by: https://github.com/eee4017, https://github.com/laithsakka ghstack dependencies: #177308

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka

pytorch#177418) ### Why - An IMA debugging aid to specifically disable CCA on a targeted block of code. - Another option is `PYTORCH_NO_CUDA_MEMORY_CACHING=1` but that is set globally. Usually I'd do this. ``` torch.cuda.caching_allocator_enable(False) try: ... finally: # make sure to clean up even on exception torch.cuda.caching_allocator_enable(True) ``` ### What Add a utility that - Disables CUDA caching allocator (CCA) when entering the block. - Restores the CCA state when exiting the block (even on exceptions). ``` with torch.cuda.caching_allocator_disabled(): ... ``` Pull Request resolved: pytorch#177418 Approved by: https://github.com/eee4017, https://github.com/laithsakka ghstack dependencies: pytorch#177308

Support negative index slicing with backed symints

1da3c72

[ghstack-poisoned]

ColinPeppler mentioned this pull request Mar 12, 2026

Slicing with backed should produce backed output when possible #175819

Closed

pytorch-bot bot added ciflow/inductor ciflow/torchtitan Run TorchTitan integration tests module: inductor labels Mar 12, 2026

ColinPeppler added a commit that referenced this pull request Mar 12, 2026

Support negative index slicing with backed symints

23804ec

ghstack-source-id: b6111aa Pull Request resolved: #177308

ColinPeppler added the topic: not user facing topic category label Mar 12, 2026

ColinPeppler requested a review from laithsakka March 12, 2026 21:21

laithsakka reviewed Mar 12, 2026

View reviewed changes

torch/_subclasses/fake_impls.py Show resolved Hide resolved

ColinPeppler mentioned this pull request Mar 12, 2026

Add ctx manager: caching_allocator_disabled to temporarily disable CCA #177354

Closed

ColinPeppler added 3 commits March 13, 2026 11:08

Update on "Support negative index slicing with backed symints"

fd04d6f

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Update on "Support negative index slicing with backed symints"

34e363c

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Update on "Support negative index slicing with backed symints"

bd190d9

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler mentioned this pull request Mar 13, 2026

Add ctx manager: caching_allocator_disabled to temporarily disable CCA #177418

Closed

Update on "Support negative index slicing with backed symints"

bd1aef1

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

laithsakka approved these changes Mar 16, 2026

View reviewed changes

Update on "Support negative index slicing with backed symints"

d380da6

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler added the ciflow/dtensor Run DTensor specific tests label Mar 19, 2026

pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Mar 19, 2026

pytorchmergebot added the merging label Mar 19, 2026

pytorchmergebot added the Merged label Mar 19, 2026

pytorchmergebot closed this in 9a7ae22 Mar 19, 2026

pytorchmergebot removed the merging label Mar 19, 2026

pytorchmergebot added Reverted ci-no-td Do not run TD on this PR labels Mar 19, 2026

pytorchmergebot reopened this Mar 19, 2026

izaitsevfb mentioned this pull request Mar 19, 2026

Add AI advisor dispatch to autorevert lambda (shadow mode) pytorch/test-infra#7839

Merged

2 tasks

ryanzhang22 pushed a commit to ryanzhang22/pytorch that referenced this pull request Mar 19, 2026

Support negative index slicing with backed symints (pytorch#177308)

5fe1b7b

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka ghstack dependencies: pytorch#175819

AaronWang04 pushed a commit to AaronWang04/pytorch that referenced this pull request Mar 31, 2026

Support negative index slicing with backed symints (pytorch#177308)

46e427e

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka ghstack dependencies: pytorch#175819

Update on "Support negative index slicing with backed symints"

cecccb9

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Update

9a9d750

[ghstack-poisoned]

pytorchmergebot pushed a commit that referenced this pull request Apr 2, 2026

Support negative index slicing with backed symints

0fe51a4

ghstack-source-id: 8f31f7c Pull Request resolved: #177308

Update on "Support negative index slicing with backed symints"

a7f75e4

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler added 3 commits April 6, 2026 09:38

Update on "Support negative index slicing with backed symints"

cc63d37

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Update on "Support negative index slicing with backed symints"

10c76a7

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Update on "Support negative index slicing with backed symints"

907a318

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ColinPeppler added a commit that referenced this pull request Apr 6, 2026

Support negative index slicing with backed symints

4ad0e10

ghstack-source-id: 130382e Pull Request resolved: #177308

pytorchmergebot closed this in 0239bc2 Apr 7, 2026

nklshy-aws pushed a commit to nklshy-aws/pytorch that referenced this pull request Apr 7, 2026

Support negative index slicing with backed symints (pytorch#177308)

1419677

Pull Request resolved: pytorch#177308 Approved by: https://github.com/laithsakka

Conversation

ColinPeppler commented Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Mar 12, 2026 • 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/177308

✅ No Failures

Uh oh!

pytorch-bot bot commented Mar 12, 2026

This PR needs a release notes: label

Uh oh!

laithsakka Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

laithsakka Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ColinPeppler commented Mar 19, 2026

Uh oh!

pytorchmergebot commented Mar 19, 2026

Merge started

Uh oh!

pytorchmergebot commented Mar 19, 2026

Uh oh!

pytorchmergebot commented Mar 30, 2026

Uh oh!

pytorchmergebot commented Mar 31, 2026

Uh oh!

ColinPeppler commented Apr 2, 2026

Uh oh!

pytorchmergebot commented Apr 2, 2026

Uh oh!

pytorchmergebot commented Apr 2, 2026

Uh oh!

pytorchmergebot commented Apr 3, 2026

Uh oh!

pytorchmergebot commented Apr 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ColinPeppler commented Mar 12, 2026 •

edited

Loading

pytorch-bot bot commented Mar 12, 2026 •

edited

Loading

This PR needs a `release notes:` label