tree b05bfae0478951462bcff525a9d8122c8966be64
parent c79fd24cfc38ba9850e1c7582bc2ed92238ed66e
author Aniket Dhok <aniket.dhok@ittiam.com> 1554271286 +0530
committer Yunqing Wang <yunqingwang@google.com> 1554741644 +0000

Optimize aom_quantize_b_adaptive module

Added AVX2 variant for aom_quantize_b_adaptive_c

For speed = 3 and 4 presets observed encode time reduction of
0.7% and 0.6%(averaged across multiple test cases).

Module gains improved by factor of ~1.82x
on average w.r.t SSE2 variant.

Change-Id: Ia403ae3547020699d59f009f8938ac13d8a4d7b6