sdf-jkl commented on PR #11460: URL: https://github.com/apache/arrow-rs/pull/11460#issuecomment-6083801014
@adriangb @HippoBaro [mimalloc](https://github.com/apache/arrow-rs/pull/11460#issuecomment-6069786064) overall seems to shows less variance than both [glibc's malloc](https://github.com/apache/arrow-rs/pull/11449#issuecomment-6067746393) and [jemalloc](https://github.com/apache/arrow-rs/pull/11449#issuecomment-6068328880) runs: Metric | Main vs itself | Jemalloc vs itself | Mimalloc vs itself -- | -- | -- | -- Median drift | 1.52% | 0.78% | 0.85% 95th percentile | 11.46% | 7.14% | 3.87% Largest drift | 55.8% | 52.3% | 6.1% Cases above 5% | 48 | 21 | 4 Cases above 10% | 18 | 10 | 0 This is not really about @HippoBaro's point about order-dependent measurement artifacts, but if it shows less drift between runs it could be an overall improvement -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
