coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2025-12-07 01:06:46 +08:00
Code Issues Packages Projects Releases Wiki Activity

nv12 stores ordered for better core2 performance. 221->211ms

Browse Source 
BUG=none
TEST=none
Review URL: http://webrtc-codereview.appspot.com/330001

git-svn-id: http://libyuv.googlecode.com/svn/trunk@108 16f28f9a-4ce2-e073-06de-1de4eb20be90
This commit is contained in:
fbarchard@google.com 2011-12-15 20:01:43 +00:00
parent 373cdbdc58
commit e7ae4e1062
2 changed files with 10 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.chromium
View File

@ -1,6 +1,6 @@
Name: libyuv Name: libyuv
URL: http://code.google.com/p/libyuv/ URL: http://code.google.com/p/libyuv/
Version: 107 Version: 108
License: BSD License: BSD
License File: LICENSE License File: LICENSE

17
source/planar_functions.cc
View File

@ -55,6 +55,7 @@ static void SplitUV_SSE2(const uint8* src_uv,
mov ecx, [esp + 4 + 16] // pix mov ecx, [esp + 4 + 16] // pix
pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff
psrlw xmm5, 8 psrlw xmm5, 8
sub edi, edx
convertloop: convertloop:
movdqa xmm0, [eax] movdqa xmm0, [eax]
@ -65,13 +66,12 @@ static void SplitUV_SSE2(const uint8* src_uv,
pand xmm0, xmm5 // even bytes pand xmm0, xmm5 // even bytes
pand xmm1, xmm5 pand xmm1, xmm5
packuswb xmm0, xmm1 packuswb xmm0, xmm1
movdqa [edx], xmm0
lea edx, [edx + 16]
psrlw xmm2, 8 // odd bytes psrlw xmm2, 8 // odd bytes
psrlw xmm3, 8 psrlw xmm3, 8
packuswb xmm2, xmm3 packuswb xmm2, xmm3
movdqa [edi], xmm2 movdqa [edx], xmm0
lea edi, [edi + 16] movdqa [edx + edi], xmm2
lea edx, [edx + 16]
sub ecx, 16 sub ecx, 16
ja convertloop ja convertloop
pop edi pop edi
@ -86,6 +86,8 @@ static void SplitUV_SSE2(const uint8* src_uv,
asm volatile ( asm volatile (
"pcmpeqb %%xmm5,%%xmm5 \n" "pcmpeqb %%xmm5,%%xmm5 \n"
"psrlw $0x8,%%xmm5 \n" "psrlw $0x8,%%xmm5 \n"
"sub %1,%2 \n"
"1: \n" "1: \n"
"movdqa (%0),%%xmm0 \n" "movdqa (%0),%%xmm0 \n"
"movdqa 0x10(%0),%%xmm1 \n" "movdqa 0x10(%0),%%xmm1 \n"
@ -95,13 +97,12 @@ static void SplitUV_SSE2(const uint8* src_uv,
"pand %%xmm5,%%xmm0 \n" "pand %%xmm5,%%xmm0 \n"
"pand %%xmm5,%%xmm1 \n" "pand %%xmm5,%%xmm1 \n"
"packuswb %%xmm1,%%xmm0 \n" "packuswb %%xmm1,%%xmm0 \n"
"movdqa %%xmm0,(%1) \n"
"lea 0x10(%1),%1 \n"
"psrlw $0x8,%%xmm2 \n" "psrlw $0x8,%%xmm2 \n"
"psrlw $0x8,%%xmm3 \n" "psrlw $0x8,%%xmm3 \n"
"packuswb %%xmm3,%%xmm2 \n" "packuswb %%xmm3,%%xmm2 \n"
"movdqa %%xmm2,(%2) \n" "movdqa %%xmm0,(%1) \n"
"lea 0x10(%2),%2 \n" "movdqa %%xmm2,(%1,%2) \n"
"lea 0x10(%1),%1 \n"
"sub $0x10,%3 \n" "sub $0x10,%3 \n"
"ja 1b \n" "ja 1b \n"
: "+r"(src_uv), // %0 : "+r"(src_uv), // %0