Hello,
we are working on a OpenSG cluster application with 4 servers and 1 client.
We are experiencing delays in the rendering nodes. It seems like the
changelist is not being transfered immediately to the rendering nodes, since
every change made to the client's scene graph takes an average of 10 seconds
(depending on the size of the geometry loaded at the initialization on the
client) to be drawn by the servers. For smaller geometry files, the delay
does not exist and the system runs smoothly. Sometimes with the larger
geometry files the servers seem to "catch up" with the client, reducing the
delay progressively until it disappears and the system also runs smoothly.
Moreover, the delay is different for all rendering nodes: one will display
the changes at a given time, then another, then the first one will change
again a little bit more, then a third node will catch up, etc, etc.
We are using multicast with a ClusterWindow on the client side, and a
ClusterServer+PassiveWindow on the servers. Has anybody experienced a
similar behavior?
We would also be interested in having a blocking mechanism between all
computers (the client and all the servers) on each frame, before the draw
routine, like a barrier but among the cluster. Does OpenSG have such a
mechanism?
Thanks!
Gabriel
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
