3 ms·
Fast Checksums with Portable SIMD
- jtsylve 4y agoI experimented with the proposed parallel data type extensions to the C++ standard library. I got impressive performance gains for calculating APFS fletcher checksums without resorting to compiler intrinsics or inline assembly. Gains were even more impressive when adding some simple loop unrolling: https://jtsylve.blog/post/2022/12/24/Blazingly-Fast-er-SIMD-Checksums https://jtsylve.blog/post/2022/12/24/Blazingly-Fast-er-SIMD-...