Closed jsdillon closed 1 year ago
Base: 97.04% // Head: 97.04% // No change to project coverage :thumbsup:
Coverage data is based on head (
babc29f
) compared to base (32bfdb8
). Patch coverage: 100.00% of modified lines in pull request are covered.
:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.
This PR speeds up even_odd_zeros_checker by about 25% by not doing a separate sum calculation for
zero_count_by_ant
which is only used to order antennas.FYI, just tested the speedup of even_odd_zeros_checker that @tyler-a-cox and @AaronParsons discussed today. Unfortunately, it makes it less reliable, because there are many more instances where the data and diff_data have the same real component than cases where both the real and imaginary components match. I’m finding baselines with ~20 matches, which is getting uncomfortably close to the size of an X-engine (making absolute bounds trickier). It may be that those antennas are also broken in some other way, but I’d rather eat the extra second and make this a clean test that clearly reports out which antennas are dropping packets or have broken X-engines.