Adapt Dynamo Tests to HPUs #145476

amathewc · 2025-01-23T10:20:41Z

This PR is a continuation of #144387 . Adapted two more files with the approach described below.

#MOTIVATION

We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at #126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices.

Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu )

#CHANGES

Create a separate class for test functions running on CUDA devices
Extend the functionality of these tests to include HPUs
Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes
Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices

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

pytorch-bot · 2025-01-23T10:20:45Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/145476

📄 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.

✅ No Failures

As of commit 8e48628 with merge base 53fc921 ():
💚 Looks good so far! There are no failures yet. 💚

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

amathewc · 2025-01-24T03:20:43Z

@ankurneog , @EikanWang : please review

amathewc · 2025-01-28T04:42:41Z

@cpuhrsch , @EikanWang : Could you review this ?

amathewc · 2025-02-05T10:23:37Z

@cpuhrsch , @EikanWang : Please review

EikanWang

LGTM

EikanWang · 2025-02-12T05:18:30Z

@yanboliang , may I know if you have any comments on this PR? cc @albanD

yanboliang · 2025-02-12T22:25:08Z

Looks good to me, please fix the test failures.

amathewc · 2025-02-13T03:36:23Z

@pytorchbot rebase

pytorchmergebot · 2025-02-13T03:37:52Z

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

pytorchmergebot · 2025-02-13T03:37:55Z

Rebase failed due to Command git -C /home/runner/work/pytorch/pytorch rebase refs/remotes/origin/viable/strict pull/145476/head returned non-zero exit code 1

Rebasing (1/2)
Auto-merging test/dynamo/test_functions.py
CONFLICT (content): Merge conflict in test/dynamo/test_functions.py
Auto-merging test/dynamo/test_misc.py
error: could not apply 0211d3bf590... Adapt Dynamo Tests to HPUs
hint: Resolve all conflicts manually, mark them as resolved with
hint: "git add/rm <conflicted_files>", then run "git rebase --continue".
hint: You can instead skip this commit: run "git rebase --skip".
hint: To abort and get back to the state before "git rebase", run "git rebase --abort".
hint: Disable this message with "git config set advice.mergeConflict false"
Could not apply 0211d3bf590... Adapt Dynamo Tests to HPUs

Raised by https://github.com/pytorch/pytorch/actions/runs/13299905929

Revert test/dynamo/test_functions.py to original due to CI issues. Will be fixing them in another PR.

amathewc · 2025-02-14T07:05:55Z

@pytorchbot rebase

pytorchmergebot · 2025-02-14T07:07:23Z

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

pytorchmergebot · 2025-02-14T07:07:25Z

Rebase failed due to Command git -C /home/runner/work/pytorch/pytorch rebase refs/remotes/origin/viable/strict pull/145476/head returned non-zero exit code 1

Rebasing (1/3)
Auto-merging test/dynamo/test_functions.py
CONFLICT (content): Merge conflict in test/dynamo/test_functions.py
Auto-merging test/dynamo/test_misc.py
error: could not apply 0211d3bf590... Adapt Dynamo Tests to HPUs
hint: Resolve all conflicts manually, mark them as resolved with
hint: "git add/rm <conflicted_files>", then run "git rebase --continue".
hint: You can instead skip this commit: run "git rebase --skip".
hint: To abort and get back to the state before "git rebase", run "git rebase --abort".
hint: Disable this message with "git config set advice.mergeConflict false"
Could not apply 0211d3bf590... Adapt Dynamo Tests to HPUs

Raised by https://github.com/pytorch/pytorch/actions/runs/13324267703

amathewc · 2025-02-14T07:13:37Z

@EikanWang , @yanboliang : Since test_functions.py was causing CI issues, I have reverted that file to its current merged version. Will be raising a separate PR for test_functions.py later. Would be helpful if you could reinitiate the CI tests and merge

amathewc · 2025-02-20T03:19:19Z

@yanboliang , @guangyey : Could you help in the review and merge of this PR?

guangyey · 2025-02-20T05:16:02Z

test/dynamo/test_functions.py

            return operator.truth(x) and bool(y)

-        opt_fn = torch.compile(fullgraph=True, dynamic=False)(fn)
+        opt_fn = torch.compile(dynamic=False)(fn)


Why fullgraph is removed?

@guangyey : Initially, we had two files in this PR . test_functions.py and test_misc.py. test_functions.py was causing CI issues - We have identified why and trying to fix that . In the meanwhile, decided to go ahead with just test_misc.py in this PR . Had reverted test_functions.py to the current merged version . So, this change ( removing fullgraph) might have come from some other PR.
If this is causing confusion, I can delete this PR and create a new one with just test_misc.py. Let me know , if you think that is a better approach.

guangyey · 2025-02-20T05:17:10Z

test/dynamo/test_functions.py

        opt_fn = torch.compile(fn, backend="eager", fullgraph=True)
        self.assertEqual(fn(torch.ones(3, 3)), opt_fn(torch.ones(3, 3)))

+    @unittest.skip("https://github.com/pytorch/pytorch/pull/146527 exposed a bug")


If this PR does not introduce this bug, we should not skip the UT in this PR.

@guangyey : Replied above in the previous comment .

anijain2305 · 2025-02-21T06:59:42Z

Does this PR need rebasing? There are many changes which are not related to this PR.

anijain2305 · 2025-02-21T07:00:19Z

test/dynamo/test_functions.py


        self.assertEqual(fn(inputs, x), opt_fn(inputs, x))

+    def test_udf_tuple(self):


Does this require a rebase?

amathewc · 2025-02-21T09:01:47Z

Closing this PR. Will be creating a new one with test_misc.py.

This PR is related to #145476 . That PR had two files (test_functions.py and test_misc.py) . test_functions was causing CI/rebase/merge issues and hence removed for now. This PR contains only test_misc.py. This is a continuation of #144387 . ## MOTIVATION We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at #126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices. Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu ) ## CHANGES Create a separate class for test functions running on CUDA devices Extend the functionality of these tests to include HPUs Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices PS: Most of these changes were initially part of #147609 , but closed that PR due to merge conflicts. The review comments were handled in this PR. Pull Request resolved: #149499 Approved by: https://github.com/EikanWang, https://github.com/desertfire, https://github.com/cyyever

This PR is related to pytorch#145476 . That PR had two files (test_functions.py and test_misc.py) . test_functions was causing CI/rebase/merge issues and hence removed for now. This PR contains only test_misc.py. This is a continuation of pytorch#144387 . ## MOTIVATION We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at pytorch#126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices. Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu ) ## CHANGES Create a separate class for test functions running on CUDA devices Extend the functionality of these tests to include HPUs Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices PS: Most of these changes were initially part of pytorch#147609 , but closed that PR due to merge conflicts. The review comments were handled in this PR. Pull Request resolved: pytorch#149499 Approved by: https://github.com/EikanWang, https://github.com/desertfire, https://github.com/cyyever

Adapt Dynamo Tests to HPUs

0211d3b

pytorch-bot bot added module: dynamo topic: not user facing topic category labels Jan 23, 2025

pytorchbot added the open source label Jan 23, 2025

Merge branch 'main' into dynamo_changes1

3df3f84

cpuhrsch requested review from albanD and yanboliang January 24, 2025 03:47

cpuhrsch added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jan 24, 2025

amathewc added 2 commits January 24, 2025 11:43

Merge branch 'pytorch:main' into dynamo_changes1

d536e90

Fix lint issues of test_functions.py

8b917f7

EikanWang approved these changes Feb 12, 2025

View reviewed changes

Fix lint issues of test_functions.py

8e48628

Revert test/dynamo/test_functions.py to original due to CI issues. Will be fixing them in another PR.

albanD requested review from anijain2305 and removed request for albanD February 14, 2025 21:53

guangyey reviewed Feb 20, 2025

View reviewed changes

anijain2305 reviewed Feb 21, 2025

View reviewed changes

amathewc closed this Feb 21, 2025

amathewc deleted the dynamo_changes1 branch February 21, 2025 09:01

amathewc mentioned this pull request Feb 21, 2025

Adapt test_misc.py to HPUs #147609

Closed

ankurneog mentioned this pull request Mar 5, 2025

[RFC] Generalize pytorch content for non-native device execution pytorch/rfcs#66

Open

amathewc mentioned this pull request Mar 19, 2025

Adapt test_misc.py for HPUs #149499

Closed


		self.assertEqual(fn(inputs, x), opt_fn(inputs, x))

		def test_udf_tuple(self):

Adapt Dynamo Tests to HPUs #145476

Adapt Dynamo Tests to HPUs #145476

Uh oh!

Conversation

amathewc commented Jan 23, 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 Jan 23, 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/145476

✅ No Failures

Uh oh!

amathewc commented Jan 24, 2025

Uh oh!

amathewc commented Jan 28, 2025

Uh oh!

amathewc commented Feb 5, 2025

Uh oh!

EikanWang left a comment

Choose a reason for hiding this comment

Uh oh!

EikanWang commented Feb 12, 2025

Uh oh!

yanboliang commented Feb 12, 2025

Uh oh!

amathewc commented Feb 13, 2025

Uh oh!

pytorchmergebot commented Feb 13, 2025

Uh oh!

pytorchmergebot commented Feb 13, 2025

Uh oh!

amathewc commented Feb 14, 2025

Uh oh!

pytorchmergebot commented Feb 14, 2025

Uh oh!

pytorchmergebot commented Feb 14, 2025

Uh oh!

amathewc commented Feb 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

amathewc commented Feb 20, 2025

Uh oh!

guangyey Feb 20, 2025

Choose a reason for hiding this comment

Uh oh!

amathewc Feb 20, 2025

Choose a reason for hiding this comment

Uh oh!

guangyey Feb 20, 2025

Choose a reason for hiding this comment

Uh oh!

amathewc Feb 20, 2025

Choose a reason for hiding this comment

Uh oh!

anijain2305 commented Feb 21, 2025

Uh oh!

anijain2305 Feb 21, 2025

Choose a reason for hiding this comment

Uh oh!

amathewc commented Feb 21, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

amathewc commented Jan 23, 2025 •

edited by pytorch-bot bot

Loading

pytorch-bot bot commented Jan 23, 2025 •

edited

Loading

amathewc commented Feb 14, 2025 •

edited

Loading