tree 4101552b14243924a8752c14ef3ef273c081ddb4
parent f9c1f5b86bc1757e3f7e11110b36f4b419ec9c36
author Aniket Dhok <aniket.dhok@ittiam.com> 1553695763 +0530
committer Yunqing Wang <yunqingwang@google.com> 1554053768 +0000

Optimize adaptive quantization modules

SSE2 variants of aom_quantize_b_adaptive and aom_quantize_b_32x32_adaptive
are optimized for improved performance.

For speed = 3 and 4 presets, observed encode time reduction of
1.07% and 0.67% (averaged across multiple test cases).

Performance improved by factor of ~1.7x

Change-Id: Icb0769fc44b3a75b18b530c3d0ab172d8a99a0a0