[x265] [PATCH Review only] asm: code for transpose4x4 routine

[murugan at multicorewareinc.com]

# HG changeset patch
# User Murugan Vairavel <murugan at multicorewareinc.com>
# Date 1384784621 -19800
#      Mon Nov 18 19:53:41 2013 +0530
# Node ID d24c22e915afd33a122326516b41eecf7e055934
# Parent  a4735d0fe4759c72a3af408a43723f219688eeb4
asm: code for transpose4x4 routine

diff -r a4735d0fe475 -r d24c22e915af source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Mon Nov 18 18:59:20 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Mon Nov 18 19:53:41 2013 +0530
@@ -545,6 +545,7 @@
         p.calcrecon[BLOCK_8x8] = x265_calcRecons8_sse2;
         p.calcresidual[BLOCK_4x4] = x265_getResidual4_sse2;
         p.calcresidual[BLOCK_8x8] = x265_getResidual8_sse2;
+        p.transpose[BLOCK_4x4] = x265_transpose4_sse2;
     }
     if (cpuMask & X265_CPU_SSSE3)
     {
diff -r a4735d0fe475 -r d24c22e915af source/common/x86/pixel-a.asm
--- a/source/common/x86/pixel-a.asm	Mon Nov 18 18:59:20 2013 +0530
+++ b/source/common/x86/pixel-a.asm	Mon Nov 18 19:53:41 2013 +0530
@@ -8340,3 +8340,25 @@
     jnz    .loop
 
 RET
+
+;-----------------------------------------------------------------
+; void transpose_4x4(pixel *dst, pixel *src, intptr_t stride)
+;-----------------------------------------------------------------
+INIT_XMM sse2
+cglobal transpose4, 3, 3, 4, dest, src, stride
+
+    movd         m0,    [r1]
+    movd         m1,    [r1 + r2]
+    movd         m2,    [r1 + 2 * r2]
+
+    lea          r1,    [r1 + 2 * r2]
+
+    movd         m3,    [r1 + r2]
+
+    punpcklbw    m0,    m1
+    punpcklbw    m2,    m3
+    punpcklwd    m0,    m2
+
+    movu         [r0],    m0
+
+RET
diff -r a4735d0fe475 -r d24c22e915af source/common/x86/pixel.h
--- a/source/common/x86/pixel.h	Mon Nov 18 18:59:20 2013 +0530
+++ b/source/common/x86/pixel.h	Mon Nov 18 19:53:41 2013 +0530
@@ -365,5 +365,6 @@
 void x265_getResidual8_sse2(pixel *fenc, pixel *pred, int16_t *residual, intptr_t stride);
 void x265_getResidual16_sse4(pixel *fenc, pixel *pred, int16_t *residual, intptr_t stride);
 void x265_getResidual32_sse4(pixel *fenc, pixel *pred, int16_t *residual, intptr_t stride);
+void x265_transpose4_sse2(pixel *dest, pixel *src, intptr_t stride);
 
 #endif // ifndef X265_I386_PIXEL_H