[x265] [PATCH Review Only] asm code for interp_4tap_vert_pp_8x4

[praveen at multicorewareinc.com]

# HG changeset patch
# User Praveen Tiwari
# Date 1382539425 -19800
# Node ID 9b6b4407d8ae91b49a23f24e00b01728a4cfcc56
# Parent  3f32eb263bf0a2c4f620fa7bff63bf118884adce
asm code for interp_4tap_vert_pp_8x4

diff -r 3f32eb263bf0 -r 9b6b4407d8ae source/common/x86/ipfilter8.asm
--- a/source/common/x86/ipfilter8.asm	Wed Oct 23 20:07:42 2013 +0530
+++ b/source/common/x86/ipfilter8.asm	Wed Oct 23 20:13:45 2013 +0530
@@ -638,6 +638,32 @@
 packuswb    m1,        m1
 %endmacro
 
+%macro FILTER_V8_W8_H3 0
+punpcklbw   m2,        m3
+punpcklbw   m7,        m0,        m1
+
+pmaddubsw   m2,        m6
+pmaddubsw   m7,        m5
+
+paddw       m2,        m7
+
+pmulhrsw    m2,        m4
+packuswb    m2,        m2
+%endmacro
+
+%macro FILTER_V8_W8_H4 0
+punpcklbw   m3,        m0
+punpcklbw   m7,        m1,        m2
+
+pmaddubsw   m3,        m6
+pmaddubsw   m7,        m5
+
+paddw       m3,        m7
+
+pmulhrsw    m3,        m4
+packuswb    m3,        m3
+%endmacro
+
 %macro FILTER_V8_W8_8x2 2
 FILTER_V8_W8 %1, %2
 movq        m0,        [r0 + 4 * r1]
@@ -647,6 +673,25 @@
 movh        [r2 + r3], m1
 %endmacro
 
+%macro FILTER_V8_W8_8x4 2
+FILTER_V8_W8_8x2 %1, %2
+;8x3
+lea         r6,        [r0 + 4 * r1]
+movq        m1,        [r6 + r1]
+
+FILTER_V8_W8_H3
+
+movh        [r2 + 2 * r3], m2
+
+;8x4
+movq        m2,        [r6 + 2 * r1]
+
+FILTER_V8_W8_H4
+
+lea         r5,        [r2 + 2 * r3]
+movh        [r5 + r3], m3
+%endmacro
+
 ;-----------------------------------------------------------------------------
 ; void interp_4tap_vert_pp_%1x%2(pixel *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
 ;-----------------------------------------------------------------------------
@@ -694,3 +739,10 @@
 FILTER_V8_W8_8x2 8, 2
 
 RET
+
+;-----------------------------------------------------------------------------
+; void interp_4tap_vert_pp_8x4(pixel *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
+;-----------------------------------------------------------------------------
+FILTER_V8_W8_8x4 8, 4
+
+RET