Reduce number of slow tests and balance parts #1569
+22
−34
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1569 +/- ##
==========================================
- Coverage 81.54% 81.53% -0.01%
==========================================
Files 230 230
Lines 53088 53088
Branches 9427 9427
==========================================
- Hits 43288 43286 -2
- Misses 7365 7366 +1
- Partials 2435 2436 +1 🚀 New features to boost your workflow:
|
ricardoV94
changed the title
ricardoV94
force-pushed
the
numba_tests_slow
branch
2 times, most recently
from
1976ee4 to
e929c31
Compare
ricardoV94
force-pushed
the
numba_tests_slow
branch
from
e929c31 to
2a97306
Compare
jessegrabowski
approved these changes
Aug 1, 2025
| @@ -11,6 +11,9 @@ | |||
| RTOL = ATOL = 1e-6 if floatX.endswith("64") else 1e-3 | |||
|
|
|||
|
|
|||
| @pytest.mark.skip( | |||
There was a problem hiding this comment.
Let's just delete it inside of marking it as skip?
Maybe keep one simple smoke test (with basic settings) to make sure something upstream hasn't somehow broken something
There was a problem hiding this comment.
I want to come back to it when we implement numba caching, it's a good case of painfully slow (while it's like 2s on jax)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
CI part split
SecondBroadcast tests
In general these broadcast tests are really wasteful, since they keep testing functionality that is shared for any Elemwise, regardless of the core_op, when testing the scalar op should suffice. I removed the 90 combinations we were running just to see if runtime broadcast would raise (it does ofc).
Also the chain from iterable for good tests was not doing anything, because we were not using the prefix in the helper function, so the dictionary keys were the same and overriding the first two iterables. My changes there don't actually reduce the number of iterations.
Numba pad tests
Numba CI are alwas the slowest and pad is at least 15% of it.
Since this is only testing OpFromGraph (there's no custom dispatch for pad), it seems unnecessary. There's nothing new added to the numba module because of pad
Related to #1433
Related to #1124
📚 Documentation preview 📚: https://pytensor--1569.org.readthedocs.build/en/1569/