[x265] [PATCH] asm: 10bpp code for transpose 32x32

[murugan at multicorewareinc.com]

# HG changeset patch
# User Murugan Vairavel <murugan at multicorewareinc.com>
# Date 1386140597 -19800
#      Wed Dec 04 12:33:17 2013 +0530
# Node ID ee1221fac033355129128ba5f847910e3ed49047
# Parent  8b73b22d90e1a0d70495e8b5f009a9c4fc37f258
asm: 10bpp code for transpose 32x32

diff -r 8b73b22d90e1 -r ee1221fac033 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Wed Dec 04 12:06:19 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Wed Dec 04 12:33:17 2013 +0530
@@ -529,6 +529,7 @@
         p.transpose[BLOCK_4x4] = x265_transpose4_sse2;
         p.transpose[BLOCK_8x8] = x265_transpose8_sse2;
         p.transpose[BLOCK_16x16] = x265_transpose16_sse2;
+        p.transpose[BLOCK_32x32] = x265_transpose32_sse2;
 
         p.ssim_4x4x2_core = x265_pixel_ssim_4x4x2_core_sse2;
         PIXEL_AVG(sse2);
diff -r 8b73b22d90e1 -r ee1221fac033 source/common/x86/pixel-util8.asm
--- a/source/common/x86/pixel-util8.asm	Wed Dec 04 12:06:19 2013 +0530
+++ b/source/common/x86/pixel-util8.asm	Wed Dec 04 12:33:17 2013 +0530
@@ -1039,8 +1039,76 @@
 ; void transpose_32x32(pixel *dst, pixel *src, intptr_t stride)
 ;-----------------------------------------------------------------
 INIT_XMM sse2
+%if HIGH_BIT_DEPTH
+cglobal transpose32, 3, 7, 4, dest, src, stride
+    add    r2,    r2
+    mov    r3,    r0
+    mov    r4,    r1
+    mov    r5,    64
+    mov    r6,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 16]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 32]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 48]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r4 + 16]
+    lea    r0,    [r6 + 8 * 64]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 8 * 64 + 16]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 8 * 64 + 32]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 8 * 64 + 48]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r4 + 32]
+    lea    r0,    [r6 + 16 * 64]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 16 * 64 + 16]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 16 * 64 + 32]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 16 * 64 + 48]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r4 + 48]
+    lea    r0,    [r6 + 24 * 64]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 24 * 64 + 16]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 24 * 64 + 32]
+    mov    r3,    r0
+    call   transpose8_internal
+    lea    r1,    [r1 - 8 + 2 * r2]
+    lea    r0,    [r6 + 24 * 64 + 48]
+    mov    r3,    r0
+    call   transpose8_internal
+%else
 cglobal transpose32, 3, 7, 8, dest, src, stride
-
     mov    r3,    r0
     mov    r4,    r1
     mov    r5,    r0
@@ -1058,7 +1126,7 @@
     lea    r0,    [r3 + 16 * 32 + 16]
     mov    r5,    r0
     call   transpose16_internal
-
+%endif
     RET
 
 ;-----------------------------------------------------------------