Hi Robert,

Its a long while since I worked with light points so off hand I can't
say what might be causing the periodically high cull load, you'll need
to dig into the relevant code and add timing to see what part is
taking the time.  With your scene graph it is probably best to write
it out to .osg after you've done the optimization passes on it to see
what the resulting scene graph is structured like.  Lots of separate
transforms is very in-efficient and since the osgUtil::Optimizer
doesn't know about light points (osgParticle is a high level library
than osgUtil) then there is no way for it to flatten them so it
disables the flatten for any objects called LightPoint.

Robert.

On Jan 17, 2008 11:47 AM, Robert Balfour <[EMAIL PROTECTED]> wrote:
> I have a model of hundreds of lightpoints on the ground, each lightpoint
> with an associated small light model (taxiway lights at an airport).
>
> Looking at this .osg model in osgviewer, the CULL time has a 2Hz pulse,
> where it jumps from approx 6ms up to 12ms about twice a second.
> Sometimes it will even get "stuck" up at 12+ms for a few seconds before
> dropping back down to a 2Hz pulse.
>
> What could potentially cause this pulsating cull time?  I have other
> sets of lightpoints/model (runway lights) that don't exhibit any
> pulsating.
>
> Also, even a 6ms cull time seems a bit expensive (the draw time is a
> reasonable <1ms).  Each of the hundreds of lightpoints and models were
> positioned with a STATIC matrix transform (which would be flattened and
> optimized out on load?), with a top DYNAMIC matrix transform over the
> whole group (to position it in real-world airport coordinates). Can this
> be re-organized to reduce the cull time?
>
> Thanks.
>
>
> Bob.
> --
> Robert E. Balfour, Ph.D.
> Exec. V.P. & CTO,  BALFOUR Technologies LLC
> 960 South Broadway, Suite 108, Hicksville NY 11801
> Phone: (516)513-0030  Fax: (516)513-0027  email: [EMAIL PROTECTED]
> _______________________________________________
> osg-users mailing list
> [email protected]
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
_______________________________________________
osg-users mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Reply via email to