tree 2ec01dd3dde406b1eda64bd79e96194c823322f3
parent f763deffb7fba3c1704cbfe1358f160e56ac7d7c
author Aniket Dhok <aniket.dhok@ittiam.com> 1553006729 +0530
committer Yunqing Wang <yunqingwang@google.com> 1553194363 +0000

Optimize aom_quantize_b_adaptive module

Added SSE2 variant for aom_quantize_b_adaptive_c

For speed = 1, 2, 3 and 4 presets observed encode time reduction of
1.2%, 1.84%, 5.4% and 2.38% (averaged across multiple test cases).

Module gains improved by factor of ~4.3x w.r.t C code.

Change-Id: Iee2d40653d2263805c6505e9eeb8659d2d5499c9