Rich-T-kid commented on PR #11918: URL: https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614965299
Re-ran the focused benchmarks after the `Packer` change: - Coalescer sparse cases were [24–53% faster on aarch64](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614592347), [21–54% faster on x86](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614631606), and [19–53% faster with x86-64-v3](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614638123). - Dense cases were within 2.5% on aarch64/default x86 and 1–8% faster with x86-64-v3. - Standalone Boolean filtering was [roughly flat on aarch64](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614659827). The 50% case was [6% faster on x86](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614858375) and [23% faster with x86-64-v3](https://github.com/apache/arrow-rs/pull/11918#issuecomment-6614584708), with the extreme cases roughly flat. The earlier default-x86 filter regression did not reproduce on the repeat run. -- 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]
