tree b22216149f4d1f89ad6bf90909b71dbd39ec143f
parent e27d7b714bd62ccec8ab450f5990357b7b7ee812
author Jonathan Wright <jonathan.wright@arm.com> 1755516251 +0100
committer James Zern <jzern@google.com> 1756325761 -0700

Optimize aom_convolve8_vert with 4-tap filters using Neon I8MM

Add a Neon I8MM implementation for aom_convolve8_vert that is
specialized for use with 4-tap filters. This is faster than the
current Armv8.0 Neon implementation.

Change-Id: Ib95268065f6108b9c5ce75c77594cb17c620cd68