-
Notifications
You must be signed in to change notification settings - Fork 769
[SYCL][E2E] Print features detected by lit #16866
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.
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
Conversation
c9c53d8
to
3151a52
Compare
This is a bit controversial as the output is noisy. Feel free to vote via thumb up/down. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
from the CI log doesn't look noisy
sycl/test-e2e/lit.cfg.py
Outdated
@@ -890,6 +890,13 @@ def open_check_file(file_name): | |||
("%clang", " " + config.dpcpp_compiler + " " + config.c_flags) | |||
) | |||
|
|||
lit_config.note( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How about we print all features only if --verbose
/ -v
is specified?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think it would be useful that much. This PR only makes sense if we enable it by default in all our CI jobs, meaning the noise will be there.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think it would be useful that much. This PR only makes sense if we enable it by default in all our CI jobs, meaning the noise will be there.
IIUC, we already pass -v
to llvm-lit
in our CI jobs: https://github.com/intel/llvm/blob/sycl/devops/actions/run-tests/e2e/action.yml#L62.
I don't mind printing features in CI jobs as they would be useful for debugging (it's not that noisy :)). My preference would be to not print features unnecessarily when we invoke llvm-lit
locally. We already print device aspects unconditionally and llvm-lit
's output (even without --verbose
) can be big when we have multiple devices in the environment.
I think this is a good change. I do understand the concern about noisy output, but the information (in particular the common aspects) is generally hard to otherwise gather and has an important impact on how the testing is done, so the noise at the beginning of the already fairly verbose LIT output seems worth it to me. |
Seems fine to me too, the location of the extra output is localized. |
4b6317d
to
9a1bccf
Compare
No description provided.