tree 31e2e0c77cbf36939f7432837f549f23ea925433
parent fc4213efc1dcb31e2fc0a173a1155cd69ca34a67
author Jonathan Wright <jonathan.wright@arm.com> 1663863937 +0100
committer James Zern <jzern@google.com> 1663962443 +0000

Optimize av1_convolve_2d_sr_horiz_12tap_neon using SDOT

Add an alternative AArch64 implementation of
av1_convolve_2d_sr_horiz_12tap_neon for targets that implement the
Armv8.4-A SDOT (signed dot-product) instruction.

The existing MLA-based implementation of
av1_convolve_2d_sr_horiz_12tap_neon is retained for use on target
CPUs that do not implement the SDOT instruction (or CPUs executing in
AArch32 mode.) The availability of the SDOT instruction is indicated
by the feature macro __ARM_FEATURE_DOTPROD.

Change-Id: I6db861f41c7763c59f7176dbcd0404f146cde395