[x265] [PATCH] asm: assembly code for pixel_sad_64x32

[chen]

>+    movhlps m1,  m0
>+    paddd   m0,  m1
>+    movd    eax, m0
>+    RET
Seems good,
only one problem, I know movhlps is from x264's code, but in Agner's documents, movhlps is slow than pshufd, I am not sure which is better here
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131030/81537174/attachment.html>