libvpx VP8/VP9 1.13 Released With Yet More AVX2 & Arm Neon Optimizations

[phoronix]

Phoronix: libvpx VP8/VP9 1.13 Released With Yet More AVX2 & Arm Neon Optimizations

Google engineers last week released libvpx 1.13 as the newest feature release to this library that serves as the reference software implementation for the VP8 and VP9 codecs. While Google engineers have already spent years tuning libvpx for maximum performance with Advanced Vector Extensions (AVX), the game is not over and v1.13 has yet more tuning there -- along with continued Arm Neon optimizations too...

https://www.phoronix.com/news/libvpx-1.13-Released

[feanor]

Looking forward to your extensive benchmarks for libvpx! In our tests, our Graviton3 system beat our similarly spec Xeon at 4k HDR/highbd encoding, however the Xeon was still faster at normal SDR encoding.
DISCLAIMER: Our company, Vectorcamp contributed quite a few of those NEON optimizations, especially for highbd, under a contract from Arm. The rest were done by Arm themselves.