Slicing with backed should produce backed output when possible by ColinPeppler · Pull Request #175819 · pytorch/pytorch

ColinPeppler · 2026-02-26T00:21:21Z

x[0:s1] where x.size(0) = s0-1 should produce Min(s1, s0-1)
Before this PR, it would produce u0.

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

Differential Revision: D98767572

[ghstack-poisoned]

pytorch-bot · 2026-02-26T00:21:25Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/175819

📄 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.

✅ You can merge normally! (2 Unrelated Failures)

As of commit 72cb53f with merge base e05e600 ():

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

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

pull / linux-jammy-py3.10-gcc11 / test (distributed, 1, 2, lf.linux.2xlarge) (gh) (trunk failure)
'Test'

UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:

inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable) (gh) (#174929)
detectron2_maskrcnn_r_50_fpn

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

pytorch-bot · 2026-02-26T00:21:28Z

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: fe64a32 Pull Request resolved: #175819

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: 0cb31ef Pull Request resolved: #175819

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: f28607c Pull Request resolved: #175819

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: 9406558 Pull Request resolved: #175819

laithsakka · 2026-03-02T23:08:07Z

torch/_subclasses/fake_impls.py

        return 0
    elif guard_or_false(index > size):
        return size
+    elif guard_or_false(index >= 0):


elif guard_or_false(index < 0):
return sym_max(index + size, 0) # wrap then clamp to [0, size]

?

laithsakka · 2026-03-02T23:17:36Z

torch/_inductor/sizevars.py

+        # Min/Max fallback: we can prove Min(a, b) <= b, but this type of Min/Max
+        # reasoning isn't handled in sympy yet. So, just evaluate the Min here.
+        for lhs, rhs in [(left, right), (right, left)]:
+            if isinstance(lhs, (sympy.Min, sympy.Max)) and rhs in lhs.args:


for lhs, rhs in [(left, right), (right, left)]: if isinstance(lhs, sympy.Min) and rhs in lhs.args: return lhs # Min(Min(a, b), b) = Min(a, b) if isinstance(lhs, sympy.Max) and rhs in lhs.args: return rhs # Min(Max(a, b), b) = b

less dependence on sympy implicit optimization?
??

torch/_inductor/lowering.py

laithsakka · 2026-03-02T23:41:04Z

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)


context manager ?

no ctx mgr yet atm. i think there should be one, left as a follow-up

I mean you shall use do
with ...
or try
catch
finally

if this test fail it will ruin other tests state

good point 👍

laithsakka

left some comments

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: a563cc6 Pull Request resolved: pytorch/pytorch#175819

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

ColinPeppler · 2026-03-12T20:34:56Z

Left the negative indexing handling in the next PR

Summary: * `x[0:s1]` where x.size(0) = `s0-1` should produce `Min(s1, s0-1)` * Before this PR, it would produce `u0`. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo imported-using-ghimport Test Plan: Imported from OSS Differential Revision: D98937973 Pulled By: ColinPeppler

facebook-github-tools · 2026-03-31T20:18:50Z

@pytorchbot merge -i

(Initiating merge automatically since Phabricator Diff has merged, merging with -i because oss signals were bypassed internally)

pytorchmergebot · 2026-03-31T20:21:12Z

Merge started

Your change will be merged while ignoring the following 2 checks: inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu), inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable)

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 · 2026-03-31T20:21:26Z

Merge failed

Reason: This PR has internal changes and must be landed via Phabricator! Please try reimporting/rexporting the PR!

Details for Dev Infra team

Raised by workflow job

facebook-github-tools · 2026-03-31T20:36:41Z

@pytorchbot merge -i

(Initiating merge automatically since Phabricator Diff has merged, merging with -i because oss signals were bypassed internally)

pytorchmergebot · 2026-03-31T20:38:56Z

Merge started

Your change will be merged while ignoring the following 2 checks: inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu), inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable)

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 · 2026-03-31T20:39:07Z

Merge failed

Reason: This PR has internal changes and must be landed via Phabricator! Please try reimporting/rexporting the PR!

Details for Dev Infra team

Raised by workflow job

facebook-github-tools · 2026-03-31T20:49:04Z

@pytorchbot merge -i

(Initiating merge automatically since Phabricator Diff has merged, merging with -i because oss signals were bypassed internally)

pytorchmergebot · 2026-03-31T20:51:22Z

Merge started

Your change will be merged while ignoring the following 2 checks: inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu), inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable)

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 · 2026-03-31T20:51:35Z

Merge failed

Reason: This PR has internal changes and must be landed via Phabricator! Please try reimporting/rexporting the PR!

Details for Dev Infra team

Raised by workflow job

facebook-github-tools · 2026-03-31T21:01:15Z

@pytorchbot merge -i

(Initiating merge automatically since Phabricator Diff has merged, merging with -i because oss signals were bypassed internally)

pytorchmergebot · 2026-03-31T21:03:29Z

Merge started

Your change will be merged while ignoring the following 2 checks: inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu), inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable)

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

izaitsevfb · 2026-04-01T19:57:35Z

@pytorchbot revert -c ghfirst -m 'reverted internally'

pytorchmergebot · 2026-04-01T19:59:50Z

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

#175819)" This reverts commit b0b4804. Reverted #175819 on behalf of https://github.com/izaitsevfb due to reverted internally ([comment](#175819 (comment)))

pytorchmergebot · 2026-04-01T20:00:08Z

@ColinPeppler your PR has been successfully reverted.
⚠️ This PR might contain internal changes
cc: @pytorch/pytorch-dev-infra

…le (#178899) Summary: Original PR: #175819 - it got reverted internally (D98767572 ) - i must reland as a new diff internal -> then export again (hence this diff) ### Summary * `x[0:s1]` where x.size(0) = `s0-1` should produce `Min(s1, s0-1)` * Before this PR, it would produce `u0`. imported-using-ghimport Test Plan: Imported from OSS Reviewed By: sevenEng Differential Revision: D98937973 Pulled By: ColinPeppler

Slicing with backed should produce backed output

c399684

[ghstack-poisoned]

pytorch-bot bot added ciflow/inductor module: inductor labels Feb 26, 2026

ColinPeppler added a commit that referenced this pull request Feb 26, 2026

Slicing with backed should produce backed output

251f4d9

ghstack-source-id: fe64a32 Pull Request resolved: #175819

ColinPeppler added the topic: not user facing topic category label Feb 26, 2026

Update on "Slicing with backed should produce backed output"

b596c21

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 Feb 26, 2026

Slicing with backed should produce backed output

686086a

ghstack-source-id: 0cb31ef Pull Request resolved: #175819

Update on "Slicing with backed should produce backed output"

61518b3

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 Feb 26, 2026

Slicing with backed should produce backed output

971a8f6

ghstack-source-id: f28607c Pull Request resolved: #175819

Update on "Slicing with backed should produce backed output"

58b42a3

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 Feb 26, 2026

Slicing with backed should produce backed output

14c6e62

ghstack-source-id: 9406558 Pull Request resolved: #175819

ColinPeppler requested review from laithsakka and pianpwk February 27, 2026 23:59

laithsakka reviewed Mar 2, 2026

View reviewed changes

torch/_inductor/lowering.py Show resolved Hide resolved

laithsakka reviewed Mar 2, 2026

View reviewed changes

laithsakka requested changes Mar 2, 2026

View reviewed changes

laithsakka requested changes Mar 5, 2026

View reviewed changes

Update on "Slicing with backed should produce backed output"

1e90836

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 12, 2026

Support negative index slicing with backed symints #177280

Closed

pytorch-bot bot added the ciflow/torchtitan Run TorchTitan integration tests label Mar 12, 2026

Update on "Slicing with backed should produce backed output"

7031c84

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

sandy-gags pushed a commit to sandy-gags/pytorch that referenced this pull request Mar 12, 2026

Slicing with backed should produce backed output

6c3983f

ghstack-source-id: a563cc6 Pull Request resolved: pytorch/pytorch#175819

Update on "Slicing with backed should produce backed output"

3f9228d

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 12, 2026

Support negative index slicing with backed symints #177308

Closed

laithsakka changed the title ~~Slicing with backed should produce backed output~~ Slicing with backed should produce backed output when possible Mar 12, 2026

ColinPeppler mentioned this pull request Mar 31, 2026

[reland] Slicing with backed should produce backed output when possible (#178899) #178899

Closed

pytorchmergebot added the merging label Mar 31, 2026

pytorchmergebot removed the merging label Mar 31, 2026

pytorchmergebot added the merging label Mar 31, 2026

pytorchmergebot removed the merging label Mar 31, 2026

pytorchmergebot added the merging label Mar 31, 2026

pytorchmergebot removed the merging label Mar 31, 2026

pytorchmergebot added the merging label Mar 31, 2026

pytorchmergebot closed this in b0b4804 Mar 31, 2026

pytorchmergebot removed the merging label Mar 31, 2026

pytorchmergebot reopened this Apr 1, 2026

ColinPeppler mentioned this pull request Apr 1, 2026

Revert D98767572: Slicing with backed should produce backed output when possible #179006

Closed

Conversation

ColinPeppler commented Feb 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Feb 26, 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/175819

✅ You can merge normally! (2 Unrelated Failures)

Uh oh!

pytorch-bot bot commented Feb 26, 2026

This PR needs a release notes: label

Uh oh!

laithsakka Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

laithsakka Mar 2, 2026 • 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!

Uh oh!

laithsakka Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

ColinPeppler 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!

ColinPeppler Mar 13, 2026

Choose a reason for hiding this comment

Uh oh!

laithsakka left a comment

Choose a reason for hiding this comment

Uh oh!

ColinPeppler commented Mar 12, 2026

Uh oh!

facebook-github-tools bot commented Mar 31, 2026

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge started

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge failed

Uh oh!

facebook-github-tools bot commented Mar 31, 2026

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge started

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge failed

Uh oh!

facebook-github-tools bot commented Mar 31, 2026

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge started

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge failed

Uh oh!

facebook-github-tools bot commented Mar 31, 2026

Uh oh!

pytorchmergebot commented Mar 31, 2026

Merge started

Uh oh!

izaitsevfb commented Apr 1, 2026

Uh oh!

pytorchmergebot commented Apr 1, 2026

Uh oh!

pytorchmergebot commented Apr 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

ColinPeppler commented Feb 26, 2026 •

edited

Loading

pytorch-bot bot commented Feb 26, 2026 •

edited

Loading

This PR needs a `release notes:` label

laithsakka Mar 2, 2026 •

edited

Loading