[x265] [PATCH] asm: assembly code for pixel_sad_16x32

[dnyaneshwar at multicorewareinc.com]

# HG changeset patch
# User Dnyaneshwar Gorade <dnyaneshwar at multicorewareinc.com>
# Date 1383026222 -19800
#      Tue Oct 29 11:27:02 2013 +0530
# Node ID 4601159b4a77737e138ea6957a017bcb7bfeccb2
# Parent  0666d56aaa4259590dffdd5aa3bc7859c4763931
asm: assembly code for pixel_sad_16x32

diff -r 0666d56aaa42 -r 4601159b4a77 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Mon Oct 28 16:13:05 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Tue Oct 29 11:27:02 2013 +0530
@@ -235,6 +235,8 @@
         //p.pixelavg_pp[LUMA_4x4]  = x265_pixel_avg_4x4_mmx2;
         //PIXEL_AVE(sse2);
 
+        p.sad[LUMA_16x32]  = x265_pixel_sad_16x32_sse2;
+
         ASSGN_SSE(sse2);
         INIT2(sad, _sse2);
         INIT2(sad_x3, _sse2);
diff -r 0666d56aaa42 -r 4601159b4a77 source/common/x86/sad-a.asm
--- a/source/common/x86/sad-a.asm	Mon Oct 28 16:13:05 2013 +0530
+++ b/source/common/x86/sad-a.asm	Tue Oct 29 11:27:02 2013 +0530
@@ -119,6 +119,25 @@
     RET
 %endmacro
 
+%macro PROCESS_SAD_16x4 0
+    movu    m1,  [r2]
+    movu    m2,  [r2 + r3]
+    psadbw  m1,  [r0]
+    psadbw  m2,  [r0 + r1]
+    paddd   m1,  m2
+    paddd   m0,  m1
+    lea     r2,  [r2 + 2 * r3]
+    lea     r0,  [r0 + 2 * r1]
+    movu    m1,  [r2]
+    movu    m2,  [r2 + r3]
+    psadbw  m1,  [r0]
+    psadbw  m2,  [r0 + r1]
+    paddd   m1,  m2
+    paddd   m0,  m1
+    lea     r2,  [r2 + 2 * r3]
+    lea     r0,  [r0 + 2 * r1]
+%endmacro
+
 %macro SAD_W16 0
 ;-----------------------------------------------------------------------------
 ; int pixel_sad_16x16( uint8_t *, intptr_t, uint8_t *, intptr_t )
@@ -223,6 +242,41 @@
     paddw   m0, m1
     paddw   m0, m3
     SAD_END_SSE2
+
+;-----------------------------------------------------------------------------
+; int pixel_sad_16x32( uint8_t *, intptr_t, uint8_t *, intptr_t )
+;-----------------------------------------------------------------------------
+cglobal pixel_sad_16x32, 4,4,3
+    pxor m0, m0
+
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+    PROCESS_SAD_16x4
+
+    movu    m1,  [r2]
+    movu    m2,  [r2 + r3]
+    psadbw  m1,  [r0]
+    psadbw  m2,  [r0 + r1]
+    paddd   m1,  m2
+    paddd   m0,  m1
+    lea     r2,  [r2 + 2 * r3]
+    lea     r0,  [r0 + 2 * r1]
+    movu    m1,  [r2]
+    movu    m2,  [r2 + r3]
+    psadbw  m1,  [r0]
+    psadbw  m2,  [r0 + r1]
+    paddd   m1,  m2
+    paddd   m0,  m1
+
+    movhlps m1, m0
+    paddd   m0, m1
+    movd    eax, m0
+    RET
+
 %endmacro
 
 INIT_XMM sse2