Doug,
> We aren't actually eliminating GC. We just try to limit it
> to certain operations. Our approach is to try to prevent GC
> unless something is being added to or removed from the scene
> graph. We do this through a number of ways like keeping free lists
> of objects and using static temporary variables. This helps with
> GC at the expense of memory and more complex locking schemes.
> We would rather not have to employ some of these techniques. But
> given the nature of realtime 3D animation/simulation/visualization
> and the state of current JVM's, we must do these things.
First, thanks for remembering to address this question.
At first reading, it seems then that most of the GC problem would
probably arise from the way the app or applet is coded. That possibly
using some of the same techniques would improve performance further.
Or is that to "simplistic"? Not that I necessarily think it's a good idea
in the long run. Just by way of an experiment I think.
Have you any sense that HotSpot will be an improvement?
Brandie
=====================================================================
To subscribe/unsubscribe, send mail to [EMAIL PROTECTED]
Java 3D Home Page: http://java.sun.com/products/java-media/3D/
