Hello Michael,
Michael Raab wrote:
> the crash was a little bit misleading. The size of the changelist is not the
> problem. It seems that the sync of geometries that contain LINE_STRIPS causes
> the crashes. I've removed all geometries of this type and our application
> runs fine both in single and in cluster mode. If I put the LINE_STRIPS back
> in, the cluster mode crashes in sync.
hm, strange. I don't see a reason why GL_LINE_STRIP would make any part
of the sync behave different than for any of the other primitive types.
Are these geometries "unusual" or different from the rest in your
application in any other way? For example, do they use no index or a
complicated interleaved index? Do they have a material that is distinct
from the other materials? etc.
> We're using the OpenSG osb loader, don't know if this is a general sync
> problem or if the loader does something strange.
Does it make a difference if you run the default graph ops or not? Using
SceneFileHandler::the()->read("filename.osb", NULL);
will prevent them from running (i.e. the second argument to read is a
GraphOpSeq*, a list of graph ops to apply to the loaded model).
Cheers,
Carsten
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
