diff options
| author | bors <bors@rust-lang.org> | 2024-01-16 09:12:10 +0000 |
|---|---|---|
| committer | bors <bors@rust-lang.org> | 2024-01-16 09:12:10 +0000 |
| commit | 18abb126a384dd295d08e1f365a6996672f65674 (patch) | |
| tree | b49f99c2d9996c633dba0127997f420d7515fa83 /tests/codegen/src-hash-algorithm | |
| parent | 5df53c96127095d75c0724f545ccd280dc0fafa9 (diff) | |
| parent | df538288e09a45483a56a9af99fb7fd5333d53d8 (diff) | |
| download | rust-18abb126a384dd295d08e1f365a6996672f65674.tar.gz rust-18abb126a384dd295d08e1f365a6996672f65674.zip | |
Auto merge of #16350 - roife:neon-support-for-line-index, r=Veykril
internal: Speedup line index calculation via NEON for aarch64 This commit provides SIMD acceleration (via NEON) for `line-index` library on aarch64 architecture, which improves performance for Apple Silicon users (and potentially for future aarch64-based chips). The algorithm used here follows the same process as the original implementation using SSE2. Most of the vector instructions in SSE2 have corresponding parts in neon. The only issue is that there is no corresponding instruction for `_mm_movemask_epi8` in neon. To address this problem, I referred to the article at https://community.arm.com/arm-community-blogs/b/infrastructure-solutions-blog/posts/porting-x86-vector-bitmask-optimizations-to-arm-neon.
Diffstat (limited to 'tests/codegen/src-hash-algorithm')
0 files changed, 0 insertions, 0 deletions