Skip to content

[Linter] Improve device-bias linter by adding detection for with torch.device("cuda").#159926

Closed
etaf wants to merge 4 commits intogh/etaf/151/basefrom
gh/etaf/151/head
Closed

[Linter] Improve device-bias linter by adding detection for with torch.device("cuda").#159926
etaf wants to merge 4 commits intogh/etaf/151/basefrom
gh/etaf/151/head

Conversation

@etaf
Copy link
Collaborator

@etaf etaf commented Aug 6, 2025 •
edited
Loading

Stack from ghstack (oldest at bottom):

For example, detect the following situation:
>>>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `@requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

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

@etaf
[Linter] Improve device-bias linter by adding test for `with
fcb145c 
torch.device("cuda")`.
For example, detect the following situation:
>>> Lint for test/dynamo/test_modes.py:

  Error (TEST_DEVICE_BIAS) [device-bias]
    `@requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

[ghstack-poisoned]
@etaf etaf mentioned this pull request Aug 6, 2025
Closed
@pytorch-bot
Copy link

pytorch-bot bot commented Aug 6, 2025 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/159926

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

✅ You can merge normally! (1 Unrelated Failure)

As of commit b7c9649 with merge base c03a734 (image):

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

  • pull / linux-jammy-py3_9-clang9-xla / test (xla, 1, 1, lf.linux.12xlarge, unstable) (gh) (#158876)
    /var/lib/jenkins/workspace/xla/torch_xla/csrc/runtime/BUILD:476:14: Compiling torch_xla/csrc/runtime/xla_util_test.cpp failed: (Exit 1): gcc failed: error executing CppCompile command (from target //torch_xla/csrc/runtime:xla_util_test) /usr/bin/gcc -U_FORTIFY_SOURCE -fstack-protector -Wall -Wunused-but-set-parameter -Wno-free-nonheap-object -fno-omit-frame-pointer -g0 -O2 '-D_FORTIFY_SOURCE=1' -DNDEBUG -ffunction-sections ... (remaining 229 arguments skipped)

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

@pytorch-bot pytorch-bot bot added the topic: not user facing topic category label Aug 6, 2025
etaf added a commit that referenced this pull request Aug 6, 2025
@etaf
[Linter] Improve device-bias linter by adding test for `with
a8a1b9f 
torch.device("cuda")`.
For example, detect the following situation:
>>> Lint for test/dynamo/test_modes.py:

  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

ghstack-source-id: 5642a05
Pull Request resolved: #159926
@etaf etaf changed the title [Linter] Improve device-bias linter by adding test for `with [Linter] Improve device-bias linter by adding test for with torch.device("cuda"). Aug 6, 2025
@etaf
Update on "[Linter] Improve device-bias linter by adding test for `wi…
fd9b187 
…th torch.device("cuda")`."


For example, detect the following situation:
\>\>\>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

[ghstack-poisoned]
etaf added a commit that referenced this pull request Aug 6, 2025
@etaf
[Linter] Improve device-bias linter by adding test for `with
41ea2ee 
torch.device("cuda")`.
For example, detect the following situation:
>>> Lint for test/dynamo/test_modes.py:

  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

ghstack-source-id: 2469c7f
Pull Request resolved: #159926
@etaf etaf changed the title [Linter] Improve device-bias linter by adding test for with torch.device("cuda"). [Linter] Improve device-bias linter by adding detection for with torch.device("cuda"). Aug 6, 2025
@etaf etaf requested a review from malfet August 6, 2025 07:35
@etaf
Update on "[Linter] Improve device-bias linter by adding detection fo…
56e3263 
…r `with torch.device("cuda")`."


For example, detect the following situation:
\>\>\>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

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

[ghstack-poisoned]
etaf added a commit that referenced this pull request Aug 6, 2025
@etaf
[Linter] Improve device-bias linter by adding test for `with
9a33977 
torch.device("cuda")`.
For example, detect the following situation:
>>> Lint for test/dynamo/test_modes.py:

  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

ghstack-source-id: fdd0c20
Pull Request resolved: #159926
@etaf etaf mentioned this pull request Aug 6, 2025
Closed
@etaf
Update on "[Linter] Improve device-bias linter by adding detection fo…
b7c9649 
…r `with torch.device("cuda")`."


For example, detect the following situation:
\>\>\>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(

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

[ghstack-poisoned]
@etaf etaf requested a review from jansel August 6, 2025 19:59
@etaf
Copy link
Collaborator Author

etaf commented Aug 8, 2025

@pytorchbot merge

@pytorch-bot pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Aug 8, 2025
@pytorchmergebot
Copy link
Collaborator

pytorchmergebot commented Aug 8, 2025

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
Copy link
Collaborator

pytorchmergebot commented Aug 8, 2025

Starting merge as part of PR stack under #159949

hinriksnaer pushed a commit to hinriksnaer/pytorch that referenced this pull request Aug 8, 2025
@etaf @hinriksnaer
[Linter] Improve device-bias linter by adding detection for `with tor…
4e99bfd 
…ch.device("cuda")`. (pytorch#159926)

```
For example, detect the following situation:
>>>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `@requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(
```

Pull Request resolved: pytorch#159926
Approved by: https://github.com/EikanWang, https://github.com/jansel
ghstack dependencies: pytorch#159759
@github-actions github-actions bot deleted the gh/etaf/151/head branch September 8, 2025 02:13
markc-614 pushed a commit to markc-614/pytorch that referenced this pull request Sep 17, 2025
@etaf @markc-614
[Linter] Improve device-bias linter by adding detection for `with tor…
a495df5 
…ch.device("cuda")`. (pytorch#159926)

```
For example, detect the following situation:
>>>Lint for test/dynamo/test_modes.py:
  Error (TEST_DEVICE_BIAS) [device-bias]
    `@requires_gpu` function should not hardcode `with torch.device('cuda')`,
    suggest to use torch.device(GPU_TYPE)

        687  |            flex_attention as flex_attention_eager,
        688  |        )
        689  |
    >>> 690  |        with torch.device("cuda"):
        691  |            flex_attention = torch.compile(flex_attention_eager, dynamic=False)
        692  |
        693  |            with self.assertRaisesRegex(
```

Pull Request resolved: pytorch#159926
Approved by: https://github.com/EikanWang, https://github.com/jansel
ghstack dependencies: pytorch#159759
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@jansel jansel jansel approved these changes

@EikanWang EikanWang EikanWang approved these changes

@malfet malfet Awaiting requested review from malfet

Assignees

No one assigned

Labels

ciflow/inductor ciflow/trunk Trigger trunk jobs on your pull request Merged module: dynamo open source topic: not user facing topic category

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@etaf @pytorchmergebot @jansel @EikanWang @pytorchbot