[x265] [PATCH] asm: assembly code for pixel_sad_24x32

[dnyaneshwar at multicorewareinc.com]

# HG changeset patch
# User Dnyaneshwar Gorade <dnyaneshwar at multicorewareinc.com>
# Date 1383136861 -19800
#      Wed Oct 30 18:11:01 2013 +0530
# Node ID 62a51fe2fcbfd76fc8476a6f714f961b3f3f23ef
# Parent  eb7d9f928ee031a108b7c77e56f1d64f123d7157
asm: assembly code for pixel_sad_24x32

diff -r eb7d9f928ee0 -r 62a51fe2fcbf source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Wed Oct 30 15:29:59 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Wed Oct 30 18:11:01 2013 +0530
@@ -252,6 +252,7 @@
         p.sad[LUMA_64x64]  = x265_pixel_sad_64x64_sse2;
 
         p.sad[LUMA_48x64]  = x265_pixel_sad_48x64_sse2;
+        p.sad[LUMA_24x32]  = x265_pixel_sad_24x32_sse2;
 
         ASSGN_SSE(sse2);
         INIT2(sad, _sse2);
diff -r eb7d9f928ee0 -r 62a51fe2fcbf source/common/x86/pixel.h
--- a/source/common/x86/pixel.h	Wed Oct 30 15:29:59 2013 +0530
+++ b/source/common/x86/pixel.h	Wed Oct 30 18:11:01 2013 +0530
@@ -52,6 +52,7 @@
     ret x265_pixel_ ## name ## _64x48_ ## suffix args; \
     ret x265_pixel_ ## name ## _64x64_ ## suffix args; \
     ret x265_pixel_ ## name ## _48x64_ ## suffix args; \
+    ret x265_pixel_ ## name ## _24x32_ ## suffix args; \
 
 #define DECL_X1(name, suffix) \
     DECL_PIXELS(int, name, suffix, (pixel *, intptr_t, pixel *, intptr_t))
diff -r eb7d9f928ee0 -r 62a51fe2fcbf source/common/x86/sad-a.asm
--- a/source/common/x86/sad-a.asm	Wed Oct 30 15:29:59 2013 +0530
+++ b/source/common/x86/sad-a.asm	Wed Oct 30 18:11:01 2013 +0530
@@ -138,6 +138,44 @@
     lea     r0,  [r0 + 2 * r1]
 %endmacro
 
+%macro PROCESS_SAD_24x4 0
+    movu        m1,  [r2]
+    movq        m2,  [r2 + 16]
+    lea         r2,  [r2 + r3]
+    movu        m3,  [r2]
+    movq        m4,  [r2 + 16]
+    psadbw      m1,  [r0]
+    psadbw      m3,  [r0 + r1]
+    paddd       m0,  m1
+    paddd       m0,  m3
+    movq        m1,  [r0 + 16]
+    lea         r0,  [r0 + r1]
+    movq        m3,  [r0 + 16]
+    punpcklqdq  m2,  m4
+    punpcklqdq  m1,  m3
+    psadbw      m2, m1
+    paddd       m0, m2
+    lea         r2,  [r2 + r3]
+    lea         r0,  [r0 + r1]
+
+    movu        m1,  [r2]
+    movq        m2,  [r2 + 16]
+    lea         r2,  [r2 + r3]
+    movu        m3,  [r2]
+    movq        m4,  [r2 + 16]
+    psadbw      m1,  [r0]
+    psadbw      m3,  [r0 + r1]
+    paddd       m0,  m1
+    paddd       m0,  m3
+    movq        m1,  [r0 + 16]
+    lea         r0,  [r0 + r1]
+    movq        m3,  [r0 + 16]
+    punpcklqdq  m2,  m4
+    punpcklqdq  m1,  m3
+    psadbw      m2, m1
+    paddd       m0, m2
+%endmacro
+
 %macro PROCESS_SAD_32x4 0
     movu    m1,  [r2]
     movu    m2,  [r2 + 16]
@@ -942,6 +980,33 @@
     movd    eax, m0
     RET
 
+;-----------------------------------------------------------------------------
+; int pixel_sad_24x32( uint8_t *, intptr_t, uint8_t *, intptr_t )
+;-----------------------------------------------------------------------------
+cglobal pixel_sad_24x32, 4,4,4
+    pxor  m0,  m0
+    mov   r4,  32
+
+.loop
+    PROCESS_SAD_24x4
+    lea         r2,  [r2 + r3]
+    lea         r0,  [r0 + r1]
+    PROCESS_SAD_24x4
+    lea         r2,  [r2 + r3]
+    lea         r0,  [r0 + r1]
+    sub   r4,  8
+    cmp   r4,  8
+jnz .loop
+    PROCESS_SAD_24x4
+    lea         r2,  [r2 + r3]
+    lea         r0,  [r0 + r1]
+    PROCESS_SAD_24x4
+
+    movhlps m1,  m0
+    paddd   m0,  m1
+    movd    eax, m0
+    RET
+
 %endmacro
 
 INIT_XMM sse2