[x265] [PATCH Review only] asm: pixelsub_ps routine for 12x16 blocks

[chen]

>+;-----------------------------------------------------------------------------
>+; void pixel_sub_ps_c_%1x%2(pixel *dest, intptr_t destride, pixel *src0, pixel *src1, intptr_t srcstride0, intptr_t srcstride1);
>+;-----------------------------------------------------------------------------
>+%macro PIXELSUB_PS_W12_H4 2
>+INIT_XMM sse4
>+cglobal pixel_sub_ps_%1x%2, 6, 7, 8, dest, deststride, src0, src1, srcstride0, srcstride1
>+
>+add    r1,     r1
>+mov    r6d,    %2/4
>+
>+.loop
>+
>+    movh         m0,    [r2]
>+    movd         m1,    [r2 + 8]
>+    movh         m2,    [r3]
>+    movd         m3,    [r3 + 8]
>+
>+    movh         m4,    [r2 + r4]
>+    movd         m5,    [r2 + r4 + 8]
>+    movh         m6,    [r3 + r5]
>+    movd         m7,    [r3 + r5 + 8]
>+
>+    punpckldq    m1,    m5
>+    punpckldq    m3,    m7
>+    pmovzxbw     m0,    m0
>+    pmovzxbw     m1,    m1
>+    pmovzxbw     m2,    m2
>+    pmovzxbw     m3,    m3
>+    pmovzxbw     m4,    m4
>+    pmovzxbw     m6,    m6

movu        m0,    [r2]
movu        m1,    [r3]
movu        m2,    [r2 + r4]
movu        m3,    [r3 + r5]
 
mova        m4, m0
mova        m5, m1
punpckhdq   m4, m2
punpckhdq   m5, m3
 
pmovzxbw    m0, m0
pmovzxbw    m1, m1
pmovzxbw    m2, m2
pmovzxbw    m3, m3
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131108/9e829f97/attachment-0001.html>