[x265] [PATCH Review only] asm: code for transpose_64x64 routine

[murugan at multicorewareinc.com]

# HG changeset patch
# User Murugan Vairavel <murugan at multicorewareinc.com>
# Date 1384952067 -19800
#      Wed Nov 20 18:24:27 2013 +0530
# Node ID df257068b5acbdb80116092cef8c5af64be41294
# Parent  2ea432a36a03a7ee7e0f788181ff35ac8db20337
asm: code for transpose_64x64 routine

diff -r 2ea432a36a03 -r df257068b5ac source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Wed Nov 20 17:51:20 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Wed Nov 20 18:24:27 2013 +0530
@@ -548,6 +548,7 @@
         p.transpose[BLOCK_8x8] = x265_transpose8_sse2;
         p.transpose[BLOCK_16x16] = x265_transpose16_sse2;
         p.transpose[BLOCK_32x32] = x265_transpose32_sse2;
+        p.transpose[BLOCK_64x64] = x265_transpose64_sse2;
     }
     if (cpuMask & X265_CPU_SSSE3)
     {
diff -r 2ea432a36a03 -r df257068b5ac source/common/x86/pixel-a.asm
--- a/source/common/x86/pixel-a.asm	Wed Nov 20 17:51:20 2013 +0530
+++ b/source/common/x86/pixel-a.asm	Wed Nov 20 18:24:27 2013 +0530
@@ -7121,6 +7121,98 @@
 
     RET
 
+;-----------------------------------------------------------------
+; void transpose_64x64(pixel *dst, pixel *src, intptr_t stride)
+;-----------------------------------------------------------------
+INIT_XMM sse2
+cglobal transpose64, 3, 7, 8, dest, src, stride
+
+    mov    r3,    r0
+    mov    r4,    r1
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 16]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 32]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 48]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+
+    lea    r1,    [r4 + 16]
+    lea    r0,    [r3 + 16 * 64]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 16 * 64 + 16]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 16 * 64 + 32]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 16 * 64 + 48]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+
+    lea    r1,    [r4 + 32]
+    lea    r0,    [r3 + 32 * 64]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 32 * 64 + 16]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 32 * 64 + 32]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 32 * 64 + 48]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+
+    lea    r1,    [r4 + 48]
+    lea    r0,    [r3 + 48 * 64]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 48 * 64 + 16]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 48 * 64 + 32]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r3 + 48 * 64 + 48]
+    mov    r5,    r0
+    mov    r6,    r1
+    TRANSPOSE_16x16 64
+
+    RET
+
 ;-----------------------------------------------------------------------------
 ; void pixel_sub_ps_c_2x4(int16_t *dest, intptr_t destride, pixel *src0, pixel *src1, intptr_t srcstride0, intptr_t srcstride1);
 ;-----------------------------------------------------------------------------
diff -r 2ea432a36a03 -r df257068b5ac source/common/x86/pixel.h
--- a/source/common/x86/pixel.h	Wed Nov 20 17:51:20 2013 +0530
+++ b/source/common/x86/pixel.h	Wed Nov 20 18:24:27 2013 +0530
@@ -368,5 +368,6 @@
 void x265_transpose8_sse2(pixel *dest, pixel *src, intptr_t stride);
 void x265_transpose16_sse2(pixel *dest, pixel *src, intptr_t stride);
 void x265_transpose32_sse2(pixel *dest, pixel *src, intptr_t stride);
+void x265_transpose64_sse2(pixel *dest, pixel *src, intptr_t stride);
 
 #endif // ifndef X265_I386_PIXEL_H