tree 12a8d3871973a88b4b5a9f2f15269fbce676d30a
parent 9ad440f5ed454fa0c680d81fe0392c8a9dbf6515
author Cheng Chen <chengchen@google.com> 1512523931 -0800
committer Cheng Chen <chengchen@google.com> 1512757581 +0000

Optimize av1_jnt_convolve_2d_copy function

With shift, convolve copy no longer needs 32-bit multiplication of
two 8-bit numbers. Thus we can implement it with sse2 instead of
sse4.

Change-Id: I63e8ba414383a24f820bad4a6c607f222ec40ec2