[x265] [PATCH Review Only] chroma_hpp[4x4] for color space i420 in avx2

[aasaipriya at multicorewareinc.com]

# HG changeset patch
# User Aasaipriya Chandran <aasaipriya at multicorewareinc.com>
# Date 1417432404 -19800
#      Mon Dec 01 16:43:24 2014 +0530
# Node ID 909d343d68b3b0e34a1239085cf6acc4f2ad2f38
# Parent  ce4394109ae51c377593164d956f6e44b8f0d3df
chroma_hpp[4x4] for color space i420 in avx2

diff -r ce4394109ae5 -r 909d343d68b3 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Wed Nov 26 11:53:40 2014 +0530
+++ b/source/common/x86/asm-primitives.cpp	Mon Dec 01 16:43:24 2014 +0530
@@ -1821,6 +1821,8 @@
         p.luma_hpp[LUMA_48x64] = x265_interp_8tap_horiz_pp_48x64_avx2;
 
         p.luma_vpp[LUMA_4x4] = x265_interp_8tap_vert_pp_4x4_avx2;
+
+        p.chroma[X265_CSP_I420].filter_hpp[CHROMA_4x4] = x265_interp_4tap_horiz_pp_4x4_avx2;
     }
 #endif // if HIGH_BIT_DEPTH
 }
diff -r ce4394109ae5 -r 909d343d68b3 source/common/x86/ipfilter8.asm
--- a/source/common/x86/ipfilter8.asm	Wed Nov 26 11:53:40 2014 +0530
+++ b/source/common/x86/ipfilter8.asm	Mon Dec 01 16:43:24 2014 +0530
@@ -1195,6 +1195,60 @@
     jnz               .loop
     RET
 
+
+INIT_YMM avx2
+cglobal interp_4tap_horiz_pp_4x4, 4,6,6
+    mov             r4d, r4m
+
+%ifdef PIC
+    lea             r5, [tab_ChromaCoeff]
+    vpbroadcastd    m0, [r5 + r4 * 4]
+%else
+    vpbroadcastd    m0, [tab_ChromaCoeff + r4 * 4]
+%endif
+
+    movu            m1, [tab_Tm]
+    vpbroadcastd    m2, [pw_1]
+
+    ; register map
+    ; m0 - interpolate coeff
+    ; m1 - shuffle order table
+    ; m2 - constant word 1
+
+    sub             r0, 1
+    ; Row 0-1
+    vbroadcasti128  m3, [r0]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m3, m1
+    vbroadcasti128  m4, [r0 + r1]                   ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m4, m1
+    vperm2f128      m3, m3, m4, 00100000b
+    pmaddubsw       m3, m0
+    pmaddwd         m3, m2
+
+    ; Row 2-3
+    lea             r0, [r0 + r1 * 2]
+    vbroadcasti128  m4, [r0]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m4, m1
+    vbroadcasti128  m5, [r0 + r1]                   ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m5, m1
+    vperm2f128      m4, m4, m5, 00100000b
+    pmaddubsw       m4, m0
+    pmaddwd         m4, m2
+
+    packssdw        m3, m4
+    pmulhrsw        m3, [pw_512]
+    vextracti128    xm4, m3, 1
+    packuswb        xm3, xm4
+
+    lea             r0, [r3 * 3]
+    movd            [r2], xm3
+    pextrd          [r2+r3], xm3, 2
+    pextrd          [r2+r3*2], xm3, 1
+    pextrd          [r2+r0], xm3, 3
+    RET
+
+
+
 ;--------------------------------------------------------------------------------------------------------------
 ; void interp_8tap_horiz_pp_%1x%2(pixel *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
 ;--------------------------------------------------------------------------------------------------------------
diff -r ce4394109ae5 -r 909d343d68b3 source/common/x86/ipfilter8.h
--- a/source/common/x86/ipfilter8.h	Wed Nov 26 11:53:40 2014 +0530
+++ b/source/common/x86/ipfilter8.h	Mon Dec 01 16:43:24 2014 +0530
@@ -573,6 +573,7 @@
     SETUP_CHROMA_SS_FUNC_DEF(16, 64, cpu);
 
 CHROMA_FILTERS(_sse4);
+CHROMA_FILTERS(_avx2);
 CHROMA_SP_FILTERS(_sse2);
 CHROMA_SP_FILTERS_SSE4(_sse4);
 CHROMA_SS_FILTERS(_sse2);