Skip to content

Conversation

@aelovikov-intel
Copy link
Contributor

First, full coverage belongs to CTS, not to unit tests. Second, implementation is heavily templated and there isn't much differences potential for many intger types or vector sizes.

This PR eliminates test cases that are unlikely to excercise the code not covered by the remaining ones. The purpose of this is to reduce the time these tests takes as they're very close to be a bottleneck on machines with lots of cores where total wall clock time is defined by the longest tests in the suite.

First, full coverage belongs to CTS, not to unit tests. Second,
implementation is heavily templated and there isn't much differences
potential for many intger types or vector sizes.

This PR eliminates test cases that are unlikely to excercise the code
not covered by the remaining ones. The purpose of this is to reduce the
time these tests takes as they're very close to be a bottleneck on
machines with lots of cores where total wall clock time is defined by
the longest tests in the suite.
@aelovikov-intel
Copy link
Contributor Author

#12953 was previous attempt to deal with huge compile time by splitting the test into four (looks like it had been even worse before the split).

@aelovikov-intel
Copy link
Contributor Author

@steffenlarsen , please review post-commit.

@aelovikov-intel aelovikov-intel merged commit c309c69 into intel:sycl Jan 10, 2025
17 of 18 checks passed
@aelovikov-intel aelovikov-intel deleted the reduce-test branch January 10, 2025 15:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants