Add OpenMP based parallelization also to qsort
Closed this issue · 4 comments
Thank you for your great work on x86-simd-sort! I noticed that on May 29, a pull request titled "Adds OpenMP based parallelization to key-value sorting" was merged, which added OpenMP-based parallelization to key-value sorting and brought significant performance improvements for large arrays.
However, it seems that qsort in the library does not yet include similar thread-level parallelization.
Adding OpenMP-based parallelization to qsort could yield significant performance improvements. Would it be possible to extend the same parallelization approach to qsort in the future?
@game-difficulty I think that should be possible. I am on travel right now, we will look into it once I am back in a few weeks.
@game-difficulty #179 provides what you requested. Still under review, we will merge it soon after that.
@game-difficulty that patch has landed. Hope that helps.
exactly what I needed . Huge thanks for implementing this