[x265] [PATCH 1 of 8] asm:intra_pred_ang4_2_sse2

[dtyx265 at gmail.com]

# HG changeset patch
# User David T Yuen <dtyx265 at gmail.com>
# Date 1427138798 25200
# Node ID 5ddfa3f9d2af9272d7c30f18014ad3dacc35cb9b
# Parent  cc496665280f9e9e4776327e9a1cc1b2eeffecbc
asm:intra_pred_ang4_2_sse2

This is backported from sse4 code and replaces c code.

64-bit

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[ 2\]"
intra_ang_4x4[ 2]	8.86x 	 134.98   	 1195.68

32-bit

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[ 2\]"
intra_ang_4x4[ 2]	9.23x 	 222.48   	 2053.30

diff -r cc496665280f -r 5ddfa3f9d2af source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Sun Mar 22 22:16:45 2015 -0400
+++ b/source/common/x86/asm-primitives.cpp	Mon Mar 23 12:26:38 2015 -0700
@@ -1196,6 +1196,8 @@
         p.cu[BLOCK_16x16].intra_pred[PLANAR_IDX] = x265_intra_pred_planar16_sse2;
         p.cu[BLOCK_32x32].intra_pred[PLANAR_IDX] = x265_intra_pred_planar32_sse2;
 
+        p.cu[BLOCK_4x4].intra_pred[2] = x265_intra_pred_ang4_2_sse2;
+
         p.cu[BLOCK_4x4].calcresidual = x265_getResidual4_sse2;
         p.cu[BLOCK_8x8].calcresidual = x265_getResidual8_sse2;
 
diff -r cc496665280f -r 5ddfa3f9d2af source/common/x86/intrapred.h
--- a/source/common/x86/intrapred.h	Sun Mar 22 22:16:45 2015 -0400
+++ b/source/common/x86/intrapred.h	Mon Mar 23 12:26:38 2015 -0700
@@ -47,6 +47,8 @@
 #define DECL_ANG(bsize, mode, cpu) \
     void x265_intra_pred_ang ## bsize ## _ ## mode ## _ ## cpu(pixel* dst, intptr_t dstStride, const pixel* srcPix, int dirMode, int bFilter);
 
+DECL_ANG(4, 2, sse2);
+
 DECL_ANG(4, 2, ssse3);
 DECL_ANG(4, 3, sse4);
 DECL_ANG(4, 4, sse4);
diff -r cc496665280f -r 5ddfa3f9d2af source/common/x86/intrapred8.asm
--- a/source/common/x86/intrapred8.asm	Sun Mar 22 22:16:45 2015 -0400
+++ b/source/common/x86/intrapred8.asm	Mon Mar 23 12:26:38 2015 -0700
@@ -1109,6 +1109,29 @@
 
 %endif ; end ARCH_X86_32
 
+;-----------------------------------------------------------------------------------------
+; void intraPredAng4(pixel* dst, intptr_t dstStride, pixel* src, int dirMode, int bFilter)
+;-----------------------------------------------------------------------------------------
+INIT_XMM sse2
+cglobal intra_pred_ang4_2, 3,5,3
+    lea         r4, [r2 + 2]
+    add         r2, 10
+    cmp         r3m, byte 34
+    cmove       r2, r4
+
+    movh        m0, [r2]
+    movd        [r0], m0
+    mova        m1, m0
+    psrldq      m1, 1
+    movd        [r0 + r1], m1
+    mova        m2, m0
+    psrldq      m2, 2
+    movd        [r0 + r1 * 2], m2
+    lea         r1, [r1 * 3]
+    psrldq      m0, 3
+    movd        [r0 + r1], m0
+    RET
+
 ;---------------------------------------------------------------------------------------------
 ; void intra_pred_dc(pixel* dst, intptr_t dstStride, pixel *srcPix, int dirMode, int bFilter)
 ;---------------------------------------------------------------------------------------------