tree c2366aaf022e4cf171ec7495052e05becbbd6d6f
parent c951719d441089b714f27a01057c34eb04ab9b3c
author Kyle Siefring <siekyleb@amazon.com> 1659467794 -0400
committer Chi Yo Tsai <chiyotsai@google.com> 1660085747 +0000

Speed up mv cost table setup

Compute in O(n) instead of O(n log(n)) (where n is the number of motion
vector values). Leads to a x15 speedup with clang (~x10 with gcc due to
not autovectorizing).

Change-Id: Ibe650a9e247d864ca3b7b237d64208580a5946bb