[x265] [PATCH Review Only] asm code for interp_4tap_vert_pp_12x16

[chen]

>+xor         r4,        r4
>+add         r4d,       %2
Why not    mov r4d, %2
 
>+punpcklbw   m4,        m2,        m3,
>+punpckhbw   m5,        m2,        m3,
m2 not need anymore, why alloc a new register m5?
 
+movu        m2,        [r0 + r1]
+movu        m3,        [r0 + 2 * r1]
reduce load operators
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131028/20fba14d/attachment.html>