Hello,
Daniel Grest wrote:
Hello,
we are using OpenSG to drive our 3 sided cave, with a stero display in
the front. Everything fine so far ;)
The fps of our 'reference model' is about 40 fps,
while our previous own implementation based on openGL alone
and a special network protocol to do the synchronisation was able to
render the same model on the same hardware with about 60fps.
I am currently searching for a reason to explain the difference,
such that i can speed up my OpenSG based implementation. Theoretically
my openSg implementation should not be slower:
The scene has only one node (an textured IndexedFaceSet)
and even if i dont change anything in the graph, the fps is about 40.
What protocol are you using? Multicast or StreamSock
Have you switched on sync to vblank in your graphics driver?
What is your framerate if you draw the model in a local window?
So my question:
I remember seeing some code in one of the ClusterServer/window
classes, that looked like a synchronisation between the servers before
they swap buffers.
Is this (still) correct? Is there a second server/server sync after
the server/client sync ? If so, is it possible to turn that off?
Currently this is not possible but it is not a big change.
If you remove the swap sync, your cluster nodes will do there swap at
different times.
I'm working on the network connections. In my tests with the new code I
got up to 1000 sync cycles/s with 16 cluster nodes.
Marcus
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
