[x265] [PATCH] Luma_hpp[16x8, 16x12, 16x16 , 16x32, 16x64] avx2 asm code: improved 1229c->800c , 1723c->1109c , 2383c->1472c, 4730c->3095 , 8763c->5914c

[aasaipriya at multicorewareinc.com]

# HG changeset patch
# User Aasaipriya Chandran <aasaipriya at multicorewareinc.com>
# Date 1416470184 -19800
#      Thu Nov 20 13:26:24 2014 +0530
# Node ID 6ac5a38e0753f7966fd69fea3238719daa423161
# Parent  5b830e31ed47c7340ba5aec504303c6788dcd630
 Luma_hpp[16x8, 16x12, 16x16 , 16x32, 16x64] avx2 asm code: improved 1229c->800c , 1723c->1109c , 2383c->1472c, 4730c->3095 , 8763c->5914c

diff -r 5b830e31ed47 -r 6ac5a38e0753 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Thu Nov 20 13:24:10 2014 +0530
+++ b/source/common/x86/asm-primitives.cpp	Thu Nov 20 13:26:24 2014 +0530
@@ -1800,6 +1800,12 @@
 #endif
     p.luma_hpp[LUMA_4x4] = x265_interp_8tap_horiz_pp_4x4_avx2;
 
+    p.luma_hpp[LUMA_16x16] = x265_interp_8tap_horiz_pp_16x16_avx2;
+    p.luma_hpp[LUMA_16x8] = x265_interp_8tap_horiz_pp_16x8_avx2;
+    p.luma_hpp[LUMA_16x12] = x265_interp_8tap_horiz_pp_16x12_avx2;
+    p.luma_hpp[LUMA_16x32] = x265_interp_8tap_horiz_pp_16x32_avx2;
+    p.luma_hpp[LUMA_16x64] = x265_interp_8tap_horiz_pp_16x64_avx2;
+
     p.luma_hpp[LUMA_64x64] = x265_interp_8tap_horiz_pp_64x64_avx2;
     p.luma_hpp[LUMA_64x32] = x265_interp_8tap_horiz_pp_64x32_avx2;
     p.luma_hpp[LUMA_64x48] = x265_interp_8tap_horiz_pp_64x48_avx2;
diff -r 5b830e31ed47 -r 6ac5a38e0753 source/common/x86/ipfilter8.asm
--- a/source/common/x86/ipfilter8.asm	Thu Nov 20 13:24:10 2014 +0530
+++ b/source/common/x86/ipfilter8.asm	Thu Nov 20 13:26:24 2014 +0530
@@ -856,7 +856,76 @@
 
 
 
-
+%macro IPFILTER_LUMA_AVX2 2
+INIT_YMM avx2
+cglobal interp_8tap_horiz_pp_%1x%2, 4,6,8
+    sub             r0, 3
+    mov             r4d, r4m
+%ifdef PIC
+    lea             r5, [tab_LumaCoeff]
+    vpbroadcastd    m0, [r5 + r4 * 8]
+    vpbroadcastd    m1, [r5 + r4 * 8 + 4]
+%else
+    vpbroadcastd    m0, [tab_LumaCoeff + r4 * 8]
+    vpbroadcastd    m1, [tab_LumaCoeff + r4 * 8 + 4]
+%endif
+    ;mova            m2, [tab_Tm]
+    movu            m3, [tab_Tm + 16]
+    vpbroadcastd    m7, [pw_1]
+
+    ; register map
+    ; m0 , m1 interpolate coeff
+    ; m2 , m2  shuffle order table
+    ; m7 - pw_1
+
+    mov             r4d,  %2/2
+.loop:
+    ; Row 0
+    vbroadcasti128  m4, [r0]                        ; [x E D C B A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m5, m4, m3
+    pshufb          m4, [tab_Tm]
+    pmaddubsw       m4, m0
+    pmaddubsw       m5, m1
+    paddw           m4, m5
+    pmaddwd         m4, m7
+    vbroadcasti128  m5, [r0 + 8]                    ; second 8 elements in Row0 
+    pshufb          m6, m5, m3
+    pshufb          m5, [tab_Tm]
+    pmaddubsw       m5, m0
+    pmaddubsw       m6, m1
+    paddw           m5, m6
+    pmaddwd         m5, m7
+    packssdw        m4, m5                          ; [17 16 15 14 07 06 05 04 13 12 11 10 03 02 01 00]
+    pmulhrsw        m4, [pw_512]
+    vbroadcasti128  m2, [r0 + r1]                        ; [x E D C B A 9 8 7 6 5 4 3 2 1 0]
+    pshufb          m5, m2, m3
+    pshufb          m2, [tab_Tm]
+    pmaddubsw       m2, m0
+    pmaddubsw       m5, m1
+    paddw           m2, m5
+    pmaddwd         m2, m7
+    vbroadcasti128  m5, [r0 + r1 + 8]                    ; second 8 elements in Row0 
+    pshufb          m6, m5, m3
+    pshufb          m5, [tab_Tm]
+    pmaddubsw       m5, m0
+    pmaddubsw       m6, m1
+    paddw           m5, m6
+    pmaddwd         m5, m7
+    packssdw        m2, m5                          ; [17 16 15 14 07 06 05 04 13 12 11 10 03 02 01 00]
+    pmulhrsw        m2, [pw_512]
+    packuswb        m4, m2
+    vpermq          m4, m4, 11011000b
+    vextracti128    xm5,  m4, 1
+    pshufd          xm4, xm4, 11011000b
+    pshufd          xm5, xm5, 11011000b
+    movu            [r2], xm4
+    movu            [r2+r3],xm5
+    lea             r0, [r0 + r1 * 2]
+    lea             r2, [r2 + r3 * 2]
+    dec             r4d
+    jnz             .loop
+    RET
+%endmacro
 
 
 
@@ -982,6 +1051,12 @@
     IPFILTER_LUMA 12, 16, pp
     IPFILTER_LUMA 4, 16, pp
 
+    IPFILTER_LUMA_AVX2 16, 16
+    IPFILTER_LUMA_AVX2 16, 8
+    IPFILTER_LUMA_AVX2 16, 32
+    IPFILTER_LUMA_AVX2 16, 64
+    IPFILTER_LUMA_AVX2 16, 12 
+
 
     IPFILTER_LUMA_64x_avx2 64 , 64
     IPFILTER_LUMA_64x_avx2 64 , 48