On 8/22/06, [EMAIL PROTECTED] <
[EMAIL PROTECTED]> wrote:
The code database pager pre compile code in osgProducer::OsgCameraGroup doesn't explictly detect frame drops caused by compiles, but it will stop compiling anymore objects once the time slice is exhusted. The time slice is dynamically computed from the remaining time in the frame and a conservative estimate "fudge" factor used to make sure that it doesn't send more objects down that it estimates. This isn't a robust system though, you can't make it robust without the OpenGL calls themselves being deterministic.
Better stats collection of what happening in the OpenGL pipeline would certainly help, but alas such support isn't widely exposed.
Robert.
On Tue, 2006-08-22 at 13:03 -0600, Chris Hanson wrote:
> ... and cause
> the draw thread to be tied up compiling MUCH longer than planned -- so much so that it
> doesn't get around to completing a frame draw when required.
Is this a detectable condition in OSG? ie dropping frames due to an
OpenGL compilation request?
The code database pager pre compile code in osgProducer::OsgCameraGroup doesn't explictly detect frame drops caused by compiles, but it will stop compiling anymore objects once the time slice is exhusted. The time slice is dynamically computed from the remaining time in the frame and a conservative estimate "fudge" factor used to make sure that it doesn't send more objects down that it estimates. This isn't a robust system though, you can't make it robust without the OpenGL calls themselves being deterministic.
Better stats collection of what happening in the OpenGL pipeline would certainly help, but alas such support isn't widely exposed.
Robert.
_______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
