[Paraview] segfault during VolumeRendering of unstructuredEnSight data in latest stable (PV 2.0.1) and CVS build (PV 2.1 - April13 2005)

[Gage, Kenneth]

Dr. Avila -

Thanks for pointing this out; I think this might indeed be the problem here - a stack trace of the core file is showing that it dies in the __Construct function in libvtkVolumeRendering.so and the two machines I've been running on both have approx 1 GB of main memory - I assume this function is allocating memory but haven't checked it out yet.  I just recompiled with the debug symbols on to get more info (I hope I'm doing this right - most of my debugging experience is with Borland and it is novice-proof) and trying a build of ParaView on our SGI workstation - it has a quite a bit more memory. The 250K cells are all tetras.

My bad for floating out my previous email to all if it is confirmed to be a memory limitation - everyone was a novice once though, right? :)

Best regards -
Ken

-----Original Message-----
From: paraview-bounces at paraview.org
[mailto:paraview-bounces at paraview.org]On Behalf Of Lisa Avila
Sent: Thursday, April 21, 2005 1:24 AM
To: Kenneth Gage; paraview at paraview.org
Subject: Re: [Paraview] segfault during VolumeRendering of
unstructuredEnSight data in latest stable (PV 2.0.1) and CVS build (PV
2.1 - April13 2005)


Hi Ken,

The Bunyk ray caster is a serious memory hog - my guess is that you are 
running out a memory. You say that you have 250K cell - but what kind are 
they? If they are not tetras, then the mesh will be tetrahedralized - 
leading to many more cells. Then A LOT of information is precomputed and 
stored per point, cell, triangle face, etc. I'm not sure why your crash 
happens in ParaView but not in raw VTK - perhaps you are just on the edge 
of running out of memory in VTK, and with ParaView's GUI it pushes you over 
the limit...?

Soon (next couple of weeks) two more options for volume rendering 
unstructured grid meshes will appear in ParaView - one based on the ZSweep 
software projection method (slightly slower than Bunyk, less of a memory 
hog), and another one hardware based (much much faster than Bunyk or 
ZSweep) Hopefully having alternate methods will solve your problem!

Lisa


At 12:16 PM 4/19/2005, Kenneth Gage wrote:
>Hello -
>
>I am trying to visualize an unstructured tet dataset with 250K cells,
>70K points using ParaView.  I can get the dataset to volume render
>with the Bunyk raycaster in the development version of VTK, but both
>versions of ParaView listed above crash with a segmentation fault.
>I've attached the trace file but it doesn't appear to provide a lot of
>information to go on. Does anyone know what might be causing the
>problem, and are there any suggestions as to what tools might allow me
>to get more information on the cause of the segfault?  Perhaps a
>favored debug utility used by the ParaView team or users?
>
>The platform is an Athlon 1GHz running Fedora Core 3.  The build
>process didn't indicate any problems.
>
>All the best -
>Ken
>--
>Kenneth Gage
>MD-PhD Student
>McGowan Institute for Regenerative Medicine
>University of Pittsburgh
>Pittsburgh, PA 15203-2155
>
>_______________________________________________
>ParaView mailing list
>ParaView at paraview.org
>http://www.paraview.org/mailman/listinfo/paraview


_______________________________________________
ParaView mailing list
ParaView at paraview.org
http://www.paraview.org/mailman/listinfo/paraview