diff options
| author | divinity76 <[email protected]> | 2024-03-12 08:21:51 +0100 |
|---|---|---|
| committer | GitHub <[email protected]> | 2024-03-12 03:21:51 -0400 |
| commit | 58bea0bcbba3629043939aa499068055dd0df017 (patch) | |
| tree | 9f2b9aba4dbeca3d7e87d1c4ad4c094f17b84afc /reference_impl | |
| parent | 5b9af1c34746e20b4596c1812b683624bdcfc152 (diff) | |
optimize neon loadu_128/storeu_128 (#384)
vld1q_u8 and vst1q_u8 has no alignment requirements.
This improves performance on Oracle Cloud's VM.Standard.A1.Flex by 1.15% on a 16*1024 input, from 13920 nanoseconds down to 13800 nanoseconds (approx)
Diffstat (limited to 'reference_impl')
0 files changed, 0 insertions, 0 deletions