[x265] [PATCH 03 of 18] asm: intra_pred_ang4_4_sse2 16-bit

[dtyx265 at gmail.com]

# HG changeset patch
# User David T Yuen <dtyx265 at gmail.com>
# Date 1428074662 25200
# Node ID 77b312333ce12f2a0f62ed06719d80b44e31fbc6
# Parent  9bde344595ee60ea3beb9d09c839b897647f556f
asm: intra_pred_ang4_4_sse2 16-bit

This is backported from sse4 code and replaces c code.

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[ 4\]"
intra_ang_4x4[ 4]	4.29x 	 475.00   	 2037.57

transposed mode 32

./test/TestBench --testbench intrapred | grep "intra_ang_4x4\[32\]"
intra_ang_4x4[32]	3.82x 	 394.97   	 1510.33

diff -r 9bde344595ee -r 77b312333ce1 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Fri Apr 03 08:18:32 2015 -0700
+++ b/source/common/x86/asm-primitives.cpp	Fri Apr 03 08:24:22 2015 -0700
@@ -884,6 +884,8 @@
 
         p.cu[BLOCK_4x4].intra_pred[2] = x265_intra_pred_ang4_2_sse2;
         p.cu[BLOCK_4x4].intra_pred[3] = x265_intra_pred_ang4_3_sse2;
+        p.cu[BLOCK_4x4].intra_pred[4] = x265_intra_pred_ang4_4_sse2;
+        p.cu[BLOCK_4x4].intra_pred[32] = x265_intra_pred_ang4_4_sse2;
         p.cu[BLOCK_4x4].intra_pred[33] = x265_intra_pred_ang4_3_sse2;
 
         p.cu[BLOCK_4x4].sse_ss = x265_pixel_ssd_ss_4x4_mmx2;
diff -r 9bde344595ee -r 77b312333ce1 source/common/x86/intrapred16.asm
--- a/source/common/x86/intrapred16.asm	Fri Apr 03 08:18:32 2015 -0700
+++ b/source/common/x86/intrapred16.asm	Fri Apr 03 08:24:22 2015 -0700
@@ -779,6 +779,31 @@
     movhps      [r0 + r1], m4
     RET
 
+cglobal intra_pred_ang4_4, 3,5,8
+    mov         r4d, 2
+    cmp         r3m, byte 32
+    mov         r3d, 18
+    cmove       r3d, r4d
+
+    movu        m0, [r2 + r3]   ; [8 7 6 5 4 3 2 1]
+    mova        m2, m0
+    psrldq      m0, 2
+    punpcklwd   m2, m0      ; [5 4 4 3 3 2 2 1]
+    mova        m3, m0
+    psrldq      m0, 2
+    punpcklwd   m3, m0      ; [6 5 5 4 4 3 3 2]
+    mova        m4, m3
+    mova        m5, m0
+    psrldq      m0, 2
+    punpcklwd   m5, m0      ; [7 6 6 5 5 4 4 3]
+
+    lea         r3, [ang_table + 18 * 16]
+    mova        m0, [r3 +  3 * 16]  ; [21]
+    mova        m1, [r3 -  8 * 16]  ; [10]
+    mova        m6, [r3 + 13 * 16]  ; [31]
+    mova        m7, [r3 +  2 * 16]  ; [20]
+    jmp         mangle(private_prefix %+ _ %+ intra_pred_ang4_3 %+ SUFFIX %+ .do_filter4x4)
+
 ;-----------------------------------------------------------------------------------
 ; void intra_pred_dc(pixel* dst, intptr_t dstStride, pixel* above, int, int filter)
 ;-----------------------------------------------------------------------------------