More questions regarding the PagedLOD. Using simply a LOD gives me great
performance (as expected), however as soon as I use PagedLOD frame rate drops
significantly (from 100+ fps to like 10-15 fps).
Using the DEBUG_INFO I get nothing strange and I also tried to use
setDoPreCompile but performa
Seems to be an OSX thingy. I did some profiling and saw massive use of
GLBeginEndAdapter::Vertex3f. Apparently there is an #ifdef _APPLE_ which sets
the DrawingPolicy for OSX to USE_VERTEX_ARRAY, which isn't what readNodeFile
does. Should it be USE_VERTEX_ARRAY as default when Linux/Win has disp
On 2/17/2010 7:09 AM, Adrian Lindberg wrote:
> More questions regarding the PagedLOD. Using simply a LOD gives me great
> performance (as expected), however as soon as I use PagedLOD frame rate drops
> significantly (from 100+ fps to like 10-15 fps).
> Using the DEBUG_INFO I get nothing strange a
Only had one, but with 4 high-poly geometries. Again, it wad due to the slow
glBegin glEnd type of drawing flagged by USE_VERTEX_ARRAY (profiler showed lots
of GLBeginEndAdapter). Didn't pick up on it initially because of readNodeFile
gives a different list (either display or VBO).
But I eventu
HI Adrian,
This issue doesn't haven't anything today with the pager. If you
osg::Geometry is dropping down to use a GLBeginEndAdapter then it's on
an OpenGL slow path which will be inefficient.
Please check you database for possible places where the osg::Geometry
in it drop everything off use of
5 matches
Mail list logo