Skip to content

Revert "[inductor] add lowering for repeat_interleave.Tensor with output size specified (#147160) (#158462)" #159798

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
wants to merge 4 commits into from
Closed

Revert "[inductor] add lowering for repeat_interleave.Tensor with output size specified (#147160) (#158462)" #159798

wants to merge 4 commits into from

Conversation

davidberard98
Copy link
Contributor

@davidberard98 davidberard98 commented Aug 4, 2025 •
edited
Loading

Stack from ghstack (oldest at bottom):

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))

which leads to errors like

/tmp/torchinductor_dberard/3h/c3hlb22fpptebupstsuhl6kexa6z3upgbnyxln7c24gfcr5747iu.py:30: unknown: block: [0,0,0], thread: [10,0,0] Assertion `index out of bounds: 0 <= tmp5 < 4` failed.

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

Differential Revision: D79591561

@davidberard98
Revert "[inductor] add lowering for repeat_interleave.Tensor with out…
545a2aa 
…put size specified (#147160) (#158462)"

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

[ghstack-poisoned]
@pytorch-bot PyTorch Bot
Copy link

pytorch-bot bot commented Aug 4, 2025 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/159798

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

❌ 2 New Failures, 81 Pending

As of commit 8f77523 with merge base fb8f32e (image):

NEW FAILURES - The following jobs have failed:

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

davidberard98 added a commit that referenced this pull request Aug 4, 2025
@davidberard98
Revert "[inductor] add lowering for repeat_interleave.Tensor with out…
6725640 
…put size specified (#147160) (#158462)"

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

ghstack-source-id: 09717d5
Pull Request resolved: #159798
@davidberard98
Copy link
Contributor Author

@davidberard98 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

@davidberard98 davidberard98 added the topic: not user facing topic category label Aug 4, 2025
@pytorch-bot pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Aug 4, 2025
@davidberard98 davidberard98 mentioned this pull request Aug 4, 2025
Closed
@davidberard98 davidberard98 reopened this Aug 4, 2025
@davidberard98
Update on "Revert "[inductor] add lowering for repeat_interleave.Tens…
066e009 
…or with output size specified (#147160) (#158462)""


This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

which leads to errors like

```
/tmp/torchinductor_dberard/3h/c3hlb22fpptebupstsuhl6kexa6z3upgbnyxln7c24gfcr5747iu.py:30: unknown: block: [0,0,0], thread: [10,0,0] Assertion `index out of bounds: 0 <= tmp5 < 4` failed.
```

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

Differential Revision: [D79591561](https://our.internmc.facebook.com/intern/diff/D79591561)

[ghstack-poisoned]
davidberard98 added a commit that referenced this pull request Aug 4, 2025
@davidberard98
Revert "[inductor] add lowering for repeat_interleave.Tensor with out…
892f119 
…put size specified (#147160) (#158462)"

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

ghstack-source-id: 1ff778e
Pull Request resolved: #159798
@davidberard98 davidberard98 removed the ci-no-td Do not run TD on this PR label Aug 4, 2025
@davidberard98
Update on "Revert "[inductor] add lowering for repeat_interleave.Tens…
b356a2d 
…or with output size specified (#147160) (#158462)""


This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

which leads to errors like

```
/tmp/torchinductor_dberard/3h/c3hlb22fpptebupstsuhl6kexa6z3upgbnyxln7c24gfcr5747iu.py:30: unknown: block: [0,0,0], thread: [10,0,0] Assertion `index out of bounds: 0 <= tmp5 < 4` failed.
```

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

Differential Revision: [D79591561](https://our.internmc.facebook.com/intern/diff/D79591561)

[ghstack-poisoned]
@pytorch-bot pytorch-bot bot added the ci-no-td Do not run TD on this PR label Aug 4, 2025
davidberard98 added a commit that referenced this pull request Aug 4, 2025
@davidberard98
Revert "[inductor] add lowering for repeat_interleave.Tensor with out…
874cca0 
…put size specified (#147160) (#158462)"

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

ghstack-source-id: 4da3b4e
Pull Request resolved: #159798
@davidberard98
Copy link
Contributor Author

@davidberard98 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

@davidberard98 davidberard98 removed the ci-no-td Do not run TD on this PR label Aug 4, 2025
@davidberard98
Copy link
Contributor Author

@pytorchbot merge -f "this is a revert; it should be relatively safe to land, and it is blocking some internal use cases that are affected by the bug"

@pytorchmergebot
Copy link
Collaborator

Merge started

Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use -f as last resort and instead consider -i/--ignore-current to continue the merge ignoring current failures. This will allow currently pending tests to finish and report signal before the merge.

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
Copy link
Collaborator

Merge failed

Reason: Command git -C /home/runner/work/pytorch/pytorch cherry-pick -x 874cca0b6395db3bf6ab3f05787eac8415540d00 returned non-zero exit code 1

Auto-merging test/inductor/test_torchinductor.py
CONFLICT (content): Merge conflict in test/inductor/test_torchinductor.py
Auto-merging test/inductor/test_torchinductor_codegen_dynamic_shapes.py
error: could not apply 874cca0b639... Revert "[inductor] add lowering for repeat_interleave.Tensor with output size specified (#147160) (#158462)"
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

@davidberard98 davidberard98 mentioned this pull request Aug 4, 2025
Closed
pytorchmergebot added a commit that referenced this pull request Aug 4, 2025
@pytorchmergebot
Revert "[mps] Turn on inductor dynamic shapes tests (#159456)"
fb8f32e 
This reverts commit 19f1f99.

Reverted #159456 on behalf of https://github.com/davidberard98 due to Sorry - this causes a merge conflict with #159798, which I'm trying to land with co-dev to resolve a sev ([comment](#159456 (comment)))
pytorchmergebot referenced this pull request Aug 4, 2025
@angelayi @pytorchmergebot
[mps] Turn on inductor dynamic shapes tests (#159456)
19f1f99 
Pull Request resolved: #159456
Approved by: https://github.com/Skylion007, https://github.com/malfet
@davidberard98
Update on "Revert "[inductor] add lowering for repeat_interleave.Tens…
8f77523 
…or with output size specified (#147160) (#158462)""


This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

which leads to errors like

```
/tmp/torchinductor_dberard/3h/c3hlb22fpptebupstsuhl6kexa6z3upgbnyxln7c24gfcr5747iu.py:30: unknown: block: [0,0,0], thread: [10,0,0] Assertion `index out of bounds: 0 <= tmp5 < 4` failed.
```

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

Differential Revision: [D79591561](https://our.internmc.facebook.com/intern/diff/D79591561)

[ghstack-poisoned]
davidberard98 added a commit that referenced this pull request Aug 4, 2025
@davidberard98
Revert "[inductor] add lowering for repeat_interleave.Tensor with out…
a395bef 
…put size specified (#147160) (#158462)"

This reverts commit 305a037.

Reason: causes device-side assertion failures when running with this repro (a minimized version of a failure seen in a real model)

```
import torch
def ri(inp, repeats, output_size):
    return torch.repeat_interleave(inp, repeats, output_size=output_size)
inp = torch.arange(0, 4, device="cuda").reshape(-1, 1)
x = torch.tensor([1, 2, 3, 4], device="cuda")
ri_c = torch.compile(ri)
print(ri(inp, x, 10))
print(ri_c(inp, x, 10))
```

ghstack-source-id: 09f0152
Pull Request resolved: #159798
@pytorch-bot pytorch-bot bot added the ci-no-td Do not run TD on this PR label Aug 4, 2025
@davidberard98
Copy link
Contributor Author

@davidberard98 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

@davidberard98
Copy link
Contributor Author

@pytorchbot merge -f "this is a revert; it should be relatively safe to land, and it is blocking some internal use cases that are affected by the bug."

@pytorchmergebot
Copy link
Collaborator

Merge started

Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use -f as last resort and instead consider -i/--ignore-current to continue the merge ignoring current failures. This will allow currently pending tests to finish and report signal before the merge.

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

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@danzimm danzimm danzimm approved these changes

Assignees
No one assigned
Labels
ci-no-td Do not run TD on this PR ciflow/inductor ciflow/trunk Trigger trunk jobs on your pull request Merged module: inductor topic: not user facing topic category
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@davidberard98 @pytorchmergebot @danzimm