-
Notifications
You must be signed in to change notification settings - Fork 113
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
Make LieAlgebraModule tests better gc-able #3913
Make LieAlgebraModule tests better gc-able #3913
Conversation
to make them gc-able earlier
for immediate gc afterwards
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #3913 +/- ##
=======================================
Coverage 83.99% 83.99%
=======================================
Files 591 591
Lines 81358 81380 +22
=======================================
+ Hits 68333 68352 +19
- Misses 13025 13028 +3
|
Thanks for the quick improvements, memory usage is still quite large, but both jobs did succeed on the first try. I did a restart of the whole job to be sure. For 1.11:
and nightly:
Previously this file had about 56 GiB of allocations when it did succeed. Maybe we should also move this test-file to the |
I'll wait for this test run to finish and then put the line into And I'll move the rewrite of the second part (to no longer need the GAP conversion) a bit up on my priority list. |
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.
thanks!
Adresses the memory issues in 1.11-nightly and nightly tests reported in #3831 (comment).
The excessive memory seems to come from the GAP iso adaption in #3831. Before this PR, the whole computation was done in GAP, the plan for the future is to not need GAP at all, but in the current state a lot of data has to be transferred (once per testcase).
If seems to help to run each testcase in its single scope to make everything garbage collectable right after the testcase ended (instead of currently keeping everything in memory until the
@testset
is over).Furthermore I disabled the longest running testcase. The correctness of everything should still be covered by the other cases.
Let's hope that all of this helps with the CI failures.