-
Notifications
You must be signed in to change notification settings - Fork 424
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
Improve Python module testing and cleanup memory leaks #26353
Improve Python module testing and cleanup memory leaks #26353
Conversation
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
bf60294
to
5c0e566
Compare
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.
Looks good generally, per our conversation the CLEARs might need a revisit in future PRs.
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
Signed-off-by: Jade Abraham <[email protected]>
@jabraham17 : Would I/we want to (and be able to) leverage the work here in the Chapel 2.3 blog release announcement to test more of the GPU code? IIRC, we disabled the testing of the first example due to its use of PyTorch? |
Very much so, all that should be required is to copy over a few test files like the pytorch .skipif, |
Make more of the Python package module tests run more frequently. While doing this, I also cleaned up as many of the Python memory leaks in the Python module as I could find.
Some of the tests rely on external Python dependencies like torch and numba. Previously, these tests were always skipped and a "compileOnly" version was tested. This PR adds a step to the skipif to install the dependencies locally if they are missing. This allows the tests to actually be executed in more cases
This PR also adds support for checking for unfree'ed references, controlled by the
--pyMemLeaks
flag. This relies on an external python dependency, so this uses similar external Python dependencies testing support as aboveTesting
start_test test/library/packages/Python
start_test -memleaks test/library/packages/Python
Future Work:
toFree
list[Reviewed by @DanilaFe]