Skip to content

arm64: refmvs: Add NEON implementation of save_tmvs

Martin Storsjö requested to merge mstorsjo/dav1d:arm-refmvs into master 
               Cortex A53       A55      A72      A73      A76  Apple M1
save_tmvs_c:     116768.4  122653.1  82587.7  90445.0  45386.8  242.1
save_tmvs_neon:   79184.7   79889.9  54720.2  54522.6  29919.6  216.4

Relative speedup compared with C:

            Cortex A53    A55    A72    A73    A76   Apple M1
save_tmvs_neon:   1.47   1.54   1.51   1.66   1.52   1.12

The second commit changes the implementation to process two blocks at a time, like the x86 implementation does. However that only gives very marginal gains on some cores, and actually makes the code slower on the other cores.

Merge request reports

VideoLAN code repository instance