Dandandan commented on PR #7650:
URL: https://github.com/apache/arrow-rs/pull/7650#issuecomment-2977784303
> I need to look into one case that reports slower now:
>
> ```
> filter: mixed_utf8, 8192, nulls: 0, selectivity: 0.8
1.23 10.4±0.21ms ? ?/sec 1.00 8.5±0.18ms ?
?/sec
> filter: mixed_utf8view (max_string_len=128), 8192, nulls: 0.1,
selectivity: 0.8 1.24 4.9±0.03ms ? ?/sec 1.00
3.9±0.02ms ? ?/sec
> ```
>
> I suspect it is something silly as the utf8 case doesn't even use the
codepath I added 🤔
Perhaps it could be impacted by the change in the other codepaths?
I think what might happen in certain cases is that one benchmark "warms" the
memory allocator by already allocating memory regions, which makes the next
benchmark go faster. Perhaps making one benchmark faster for something (doing
fewer allocations) will make others that relied on this before run slower.
What happens when running only those benchmarks without running the others?
--
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: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
