[x265] [PATCH Review only] asm: pixelsub_ps routine for 6x8 block

[chen]

>+    movlps    [r0],                 m0
>+    pextrd    [r0 + 8],             m0,    2
>+    movlps    [r0 + r1],            m2
>+    pextrd    [r0 + r1 + 8],        m2,    2
>+    movlps    [r0 + 2* r1],         m4
>+    pextrd    [r0 + 2 * r1 + 8],    m4,    2
movh is better, movlps is float path, but pextrd is integer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131111/a22594c8/attachment.html>