[x265] [PATCH 02 of 11] asm: intra_pred_ang4_26_sse2

[dtyx265 at gmail.com]

# HG changeset patch
# User David T Yuen <dtyx265 at gmail.com>
# Date 1427504802 25200
# Node ID c5fa8f49bf15818007fe6ab25eb6aec65bb203cb
# Parent  d7fee58a8466d6862e72cced5a4aa12e37294d4e
asm: intra_pred_ang4_26_sse2

This is backported from sse4 code and replaces c code.

64-bit

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[26\]"
intra_ang_4x4[26]	3.87x 	 217.84   	 843.32

32-bit

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[26\]"
intra_ang_4x4[26]	2.82x 	 320.03   	 903.85

diff -r d7fee58a8466 -r c5fa8f49bf15 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Fri Mar 27 17:51:06 2015 -0700
+++ b/source/common/x86/asm-primitives.cpp	Fri Mar 27 18:06:42 2015 -0700
@@ -1205,6 +1205,7 @@
         p.cu[BLOCK_4x4].intra_pred[8] = x265_intra_pred_ang4_8_sse2;
         p.cu[BLOCK_4x4].intra_pred[9] = x265_intra_pred_ang4_9_sse2;
         p.cu[BLOCK_4x4].intra_pred[10] = x265_intra_pred_ang4_10_sse2;
+        p.cu[BLOCK_4x4].intra_pred[26] = x265_intra_pred_ang4_26_sse2;
 
         p.cu[BLOCK_4x4].calcresidual = x265_getResidual4_sse2;
         p.cu[BLOCK_8x8].calcresidual = x265_getResidual8_sse2;
diff -r d7fee58a8466 -r c5fa8f49bf15 source/common/x86/intrapred.h
--- a/source/common/x86/intrapred.h	Fri Mar 27 17:51:06 2015 -0700
+++ b/source/common/x86/intrapred.h	Fri Mar 27 18:06:42 2015 -0700
@@ -56,6 +56,7 @@
 DECL_ANG(4, 8, sse2);
 DECL_ANG(4, 9, sse2);
 DECL_ANG(4, 10, sse2);
+DECL_ANG(4, 26, sse2);
 
 DECL_ANG(4, 2, ssse3);
 DECL_ANG(4, 3, sse4);
diff -r d7fee58a8466 -r c5fa8f49bf15 source/common/x86/intrapred8.asm
--- a/source/common/x86/intrapred8.asm	Fri Mar 27 17:51:06 2015 -0700
+++ b/source/common/x86/intrapred8.asm	Fri Mar 27 18:06:42 2015 -0700
@@ -1512,6 +1512,46 @@
     movd        [r0], m0
     RET
 
+INIT_XMM sse2
+cglobal intra_pred_ang4_26, 5,5,4
+    movd        m0, [r2 + 1]            ; [8 7 6 5 4 3 2 1]
+
+    ; store
+    movd        [r0], m0
+    movd        [r0 + r1], m0
+    movd        [r0 + r1 * 2], m0
+    lea         r3, [r1 * 3]
+    movd        [r0 + r3], m0
+
+    ; filter
+    cmp         r4m, byte 0
+    jz         .quit
+
+    pxor        m3, m3
+    punpcklbw   m0, m3
+    pshuflw     m0, m0, 0x00
+    movd        m2, [r2]
+    punpcklbw   m2, m3
+    pshuflw     m2, m2, 0x00
+    movd        m1, [r2 + 9]
+    punpcklbw   m1, m3
+    psubw       m1, m2
+    psraw       m1, 1
+    paddw       m0, m1
+    packuswb    m0, m0
+
+    movd        r2, m0
+    mov         [r0], r2b
+    shr         r2, 8
+    mov         [r0 + r1], r2b
+    shr         r2, 8
+    mov         [r0 + r1 * 2], r2b
+    shr         r2, 8
+    mov         [r0 + r3], r2b
+
+.quit:
+    RET
+
 ;---------------------------------------------------------------------------------------------
 ; void intra_pred_dc(pixel* dst, intptr_t dstStride, pixel *srcPix, int dirMode, int bFilter)
 ;---------------------------------------------------------------------------------------------