[x265] [PATCH] asm: chroma_hpp[8x4, 8x16, 8x32] for i420 avx2 - improved 289c->220c, 928c->618c, 1802c->1128c

[aasaipriya at multicorewareinc.com]

# HG changeset patch
# User Aasaipriya Chandran <aasaipriya at multicorewareinc.com>
# Date 1426482803 -19800
#      Mon Mar 16 10:43:23 2015 +0530
# Node ID 97a441776465d47d57e809dd18889c8714deeef6
# Parent  3c310c18cc2026f464c7ed1fa4369e1df6105843
asm: chroma_hpp[8x4, 8x16, 8x32] for i420 avx2 - improved 289c->220c, 928c->618c, 1802c->1128c

diff -r 3c310c18cc20 -r 97a441776465 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Mon Mar 16 10:36:29 2015 +0530
+++ b/source/common/x86/asm-primitives.cpp	Mon Mar 16 10:43:23 2015 +0530
@@ -1587,6 +1587,10 @@
         p.chroma[X265_CSP_I420].pu[CHROMA_420_32x24].filter_hpp = x265_interp_4tap_horiz_pp_32x24_avx2;
         p.chroma[X265_CSP_I420].pu[CHROMA_420_32x8].filter_hpp = x265_interp_4tap_horiz_pp_32x8_avx2;
 
+        p.chroma[X265_CSP_I420].pu[CHROMA_420_8x4].filter_hpp = x265_interp_4tap_horiz_pp_8x4_avx2;
+        p.chroma[X265_CSP_I420].pu[CHROMA_420_8x16].filter_hpp = x265_interp_4tap_horiz_pp_8x16_avx2;
+        p.chroma[X265_CSP_I420].pu[CHROMA_420_8x32].filter_hpp = x265_interp_4tap_horiz_pp_8x32_avx2;
+
         p.chroma[X265_CSP_I420].pu[CHROMA_420_32x32].filter_hps = x265_interp_4tap_horiz_ps_32x32_avx2;
         p.chroma[X265_CSP_I420].pu[CHROMA_420_16x16].filter_hps = x265_interp_4tap_horiz_ps_16x16_avx2;
         p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].filter_hps = x265_interp_4tap_horiz_ps_4x4_avx2;
diff -r 3c310c18cc20 -r 97a441776465 source/common/x86/ipfilter8.asm
--- a/source/common/x86/ipfilter8.asm	Mon Mar 16 10:36:29 2015 +0530
+++ b/source/common/x86/ipfilter8.asm	Mon Mar 16 10:43:23 2015 +0530
@@ -16790,3 +16790,79 @@
 IPFILTER_CHROMA_PP_32xN_AVX2 32, 16
 IPFILTER_CHROMA_PP_32xN_AVX2 32, 24
 IPFILTER_CHROMA_PP_32xN_AVX2 32, 8
+
+;-------------------------------------------------------------------------------------------------------------
+; void interp_4tap_horiz_pp_8xN(pixel *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx
+;-------------------------------------------------------------------------------------------------------------
+%macro IPFILTER_CHROMA_PP_8xN_AVX2 2
+INIT_YMM avx2
+cglobal interp_4tap_horiz_pp_%1x%2, 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
+    mov               r4d,          %2
+
+.loop:
+    sub               r4d,          4
+    ; Row 0
+    vbroadcasti128    m3,           [r0]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb            m3,           m1
+    pmaddubsw         m3,           m0
+    pmaddwd           m3,           m2
+
+    ; Row 1
+    vbroadcasti128    m4,           [r0 + r1]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb            m4,           m1
+    pmaddubsw         m4,           m0
+    pmaddwd           m4,           m2
+    packssdw          m3,           m4
+    pmulhrsw          m3,           [pw_512]
+    lea               r0,           [r0 + r1 * 2]
+
+    ; Row 2
+    vbroadcasti128    m4,           [r0 ]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb            m4,           m1
+    pmaddubsw         m4,           m0
+    pmaddwd           m4,           m2
+
+    ; Row 3
+    vbroadcasti128    m5,           [r0 + r1]                        ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
+    pshufb            m5,           m1
+    pmaddubsw         m5,           m0
+    pmaddwd           m5,           m2
+    packssdw          m4,           m5
+    pmulhrsw          m4,           [pw_512]
+
+    packuswb          m3,           m4
+    mova              m5,           [interp_4tap_8x8_horiz_shuf]
+    vpermd            m3,           m5,     m3
+    vextracti128      xm4,          m3,     1
+    movq              [r2],         xm3
+    movhps            [r2 + r3],    xm3
+    lea               r2,           [r2 + r3 * 2]
+    movq              [r2],         xm4
+    movhps            [r2 + r3],    xm4
+    lea               r2,           [r2 + r3 * 2]
+    lea               r0,           [r0 + r1*2]
+    test              r4d,          r4d
+    jnz               .loop
+    RET
+%endmacro
+
+IPFILTER_CHROMA_PP_8xN_AVX2   8 , 16
+IPFILTER_CHROMA_PP_8xN_AVX2   8 , 32
+IPFILTER_CHROMA_PP_8xN_AVX2   8 , 4