tree 48dcf8582ee4d46f50e6fef3aa83e94a560c9051
parent 87cab6fa6e25274e820173cc8695db38d23a3264
author Jonathan Wright <jonathan.wright@arm.com> 1659514888 +0100
committer James Zern <jzern@google.com> 1662760298 +0000

Optimize Neon temporal filter using UDOT instruction

Add an alternative AArch64 implementation of
av1_apply_temporal_filter_neon that uses the Armv8.4-A UDOT (unsigned
dot-product) instruction to implement the 5x5 SSE sum calculation.

The existing implementation is retained for use on target CPUs that do
not implement the UDOT instruction (or CPUs executing in AArch32
mode.) The availability of the UDOT instruction is indicated by the
feature macro __ARM_FEATURE_DOTPROD.

Change-Id: I4e5eb8edeb0a3b629a9aabde21058ba87afb501b