[x265] [PATCH Review only] asm: pixel_add_pp routine for 8x4 block size

[murugan at multicorewareinc.com]

# HG changeset patch
# User Murugan Vairavel <murugan at multicorewareinc.com>
# Date 1383653218 -19800
#      Tue Nov 05 17:36:58 2013 +0530
# Node ID 1f3de94bee30eebab7f10df37e674fa7360803b1
# Parent  8b4fe169680fba027b67d441fe9248a7b5694bd8
asm: pixel_add_pp routine for 8x4 block size

diff -r 8b4fe169680f -r 1f3de94bee30 source/common/x86/pixel-add8.asm
--- a/source/common/x86/pixel-add8.asm	Tue Nov 05 17:33:07 2013 +0530
+++ b/source/common/x86/pixel-add8.asm	Tue Nov 05 17:36:58 2013 +0530
@@ -52,3 +52,41 @@
 
 RET
 
+;-----------------------------------------------------------------------------
+; void pixel_add_pp_c_8x4(pixel *dest, intptr_t destride, pixel *src0, pixel *src1, intptr_t srcstride0, intptr_t srcstride1);
+;-----------------------------------------------------------------------------
+INIT_XMM sse2
+cglobal pixel_add_pp_8x4, 4, 6, 8, dest, deststride, src0, src1
+
+mov        r4d,        r4m
+mov        r5d,        r5m
+
+movh       m0,    [r2]
+movh       m1,    [r3]
+
+movh       m2,    [r2 + r4]
+movh       m3,    [r3 + r5]
+
+movh       m4,    [r2 + 2 * r4]
+movh       m5,    [r3 + 2 * r5]
+
+lea        r2,    [r2 + 2 * r4]
+lea        r3,    [r3 + 2 * r5]
+
+movh       m6,    [r2 +  r4]
+movh       m7,    [r3 +  r5]
+
+paddusb    m0,    m1
+paddusb    m2,    m3
+paddusb    m4,    m5
+paddusb    m6,    m7
+
+movh    [r0],             m0
+movh    [r0 + r1],        m2
+movh    [r0 + 2 * r1],    m4
+
+lea     r0,               [r0 + 2 * r1]
+movh    [r0 + r1],        m6
+
+RET
+