Aren
/
haiku
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

The Y check can be moved out of the loop of course, why didn't I see this 

before?! Is about 4% faster than before. If anyone sees a way to make it faster
yet, please shoot! I can watch movies fullscreen on a 2 GHz Core 2 Duo in
VESA with bilinear scaling, but it would be nice to use less CPU... :-)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26667 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
Stephan Aßmus 2008-07-28 21:30:26 +00:00
parent 2011fa73d0
commit e196501316
1 changed files with 9 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/servers/app/drawing/Painter/Painter.cpp
View File

@ -1713,9 +1713,9 @@ Painter::_DrawBitmapBilinearCopy32(agg::rendering_buffer& srcBuffer,
// buffer handle for destination to be incremented per pixel
register uint8* d = dst;
for (int32 x = xIndexL; x <= xIndexR; x++) {
const uint8* s = src + xWeights[x].index;
if (wTop == 255) {
if (wTop == 255) {
for (int32 x = xIndexL; x <= xIndexR; x++) {
const uint8* s = src + xWeights[x].index;
// This case is important to prevent out
// of bounds access at bottom edge of the source
// bitmap. If the scale is low and integer, it will
@ -1733,7 +1733,11 @@ Painter::_DrawBitmapBilinearCopy32(agg::rendering_buffer& srcBuffer,
d[1] = (s[1] * wLeft + s[5] * wRight) >> 8;
d[2] = (s[2] * wLeft + s[6] * wRight) >> 8;
}
} else {
d += 4;
}
} else {
for (int32 x = xIndexL; x <= xIndexR; x++) {
const uint8* s = src + xWeights[x].index;
if (xWeights[x].weight == 255) {
// Prevent out of bounds access on the right edge
// or simply speed up.
@ -1761,8 +1765,8 @@ Painter::_DrawBitmapBilinearCopy32(agg::rendering_buffer& srcBuffer,
d[1] = t1 >> 16;
d[2] = t2 >> 16;
}
d += 4;
}
d += 4;
}
dst += dstBPR;
}