jhorstmann commented on issue #136: URL: https://github.com/apache/arrow-rs/issues/136#issuecomment-1165674091
Related paper by google: https://opensource.googleblog.com/2022/06/Vectorized%20and%20performance%20portable%20Quicksort.html According to their paper they get a huge performance improvement with the use of sorting networks for sorting smaller partitions and by using AVX512 specific instructions for partitioning. The mentioned speedup of 9x-19x times is compared to the C++ standard library, the rust standard `sort_unstable` already has some additional optimizations so that the speedup is probably lower. Our main usecase is also sorting to indices, it's not fully clear how well that use case is supported by their algorithm. We probably don't want to maintain such an implementation inside arrow-rs, but if someone ports the algorithm to rust as a crate we should be able to use it. -- 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]
