tustvold commented on PR #2278: URL: https://github.com/apache/arrow-rs/pull/2278#issuecomment-1203092821
> BTW I think it's possible to apply SIMD on this code path to further improve the performance The motivation for the manual unrolling of the loops is so that LLVM does this for us. If you check the godbolt output linked in the PR you'll see it is using AVX2 instructions. This fits with what we've seen in general with the arrow compute kernels, where only AVX512 seems to need manual implementation. -- 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]
