Skip to content

[TEST] Skip testing triangular_solve_batched on non-default CUDA stream#26115

Closed
vishwakftw wants to merge 1 commit intopytorch:masterfrom
vishwakftw:skip-trtrs-batched-nondefault-stream
Closed

[TEST] Skip testing triangular_solve_batched on non-default CUDA stream#26115
vishwakftw wants to merge 1 commit intopytorch:masterfrom
vishwakftw:skip-trtrs-batched-nondefault-stream

Conversation

@vishwakftw
Copy link
Copy Markdown
Contributor

@vishwakftw vishwakftw commented Sep 12, 2019

This is for testing purposes.

@pytorchbot pytorchbot added the module: cuda Related to torch.cuda, and CUDA support in general label Sep 12, 2019
@vishwakftw
Copy link
Copy Markdown
Contributor Author

@zou3519, it seems like non-default CUDA stream testing caused this to fail.

@zou3519
Copy link
Copy Markdown
Contributor

zou3519 commented Sep 13, 2019

Sure, let's get this merged and file an issue for triangular_solve not working on non-default stream. Thank you for investigating!

edit: sorry, do you think that the "testing on non-default stream" is broken, or that triangular solve gives in correct behavior doesn't support non-default stream?

Comment thread test/test_cuda.py

@unittest.skipIf(not TEST_MAGMA, "no MAGMA library detected")
@unittest.skip("Spuriously failing")
@skipCUDANonDefaultStreamIf(True)
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we're going to skip this we should put a message as to why (i.e., tringular_solve doesn't work with non default stream, or there is something wrong with the non-default stream testing) and link to an issue.

Copy link
Copy Markdown
Contributor Author

@vishwakftw vishwakftw Sep 13, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe triangular_solve doesn't work with non-default stream. This is an issue that takes this generally: #21821.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

test_triangular_solve is also skipped in an analogous manner.

@cpuhrsch cpuhrsch added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Sep 17, 2019
@zou3519
Copy link
Copy Markdown
Contributor

zou3519 commented Sep 17, 2019

@vishwakftw just checking, do you want this merged (so that we can prevent regressions) or are you still using this as a testing ground?

@vishwakftw
Copy link
Copy Markdown
Contributor Author

This can be merged, but I think that @mruberry’s PR on test refactor may have superseded this.

@mruberry
Copy link
Copy Markdown
Collaborator

You can merge as is.

@zou3519
Copy link
Copy Markdown
Contributor

zou3519 commented Sep 17, 2019

Cool, I'll merge this as is. Thanks for the discussion!

Copy link
Copy Markdown
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@zou3519 is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@facebook-github-bot
Copy link
Copy Markdown
Contributor

@zou3519 merged this pull request in be97641.

@vishwakftw vishwakftw deleted the skip-trtrs-batched-nondefault-stream branch September 22, 2019 17:54
laurentdupin pushed a commit to laurentdupin/pytorch that referenced this pull request Apr 24, 2026
…orch#26115)

Summary:
This is for testing purposes.
Pull Request resolved: pytorch#26115

Differential Revision: D17433122

Pulled By: zou3519

fbshipit-source-id: bf41327e6141e9ae589fcf18254c2a8cdd868dd7
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Merged module: cuda Related to torch.cuda, and CUDA support in general open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants