[vtk-developers] VTK/Examples/Infovis/Cxx/ParallelBFS.cxx

[William McLendon]

Howdy,

It's been a while since I've played with the PBGL filters but it looks
like it runs fine on my system (you might want to make the graph
smaller, I set the lines to this to speed up rendering:


        source->SetNumberOfVertices(100);
        source->SetNumberOfEdges(800);

One thing I am seeing on my system is that if I try to run it through
mpirun then I get an error because I don't have things set up for
x-forwarding under MPI.  Probably there's a setting I need to set to
allow the shells to do the right thing.  

Have you enabled testing?  If you run ctest -R "PBGL" it should run all
the PBGL tests if they are built.   I'm seeing something like this on my
system:

        
        ParaView3-build] > ctest -R "PBGL"
        Start processing tests
        Test project /home/wcmclen/dev/Titan/ParaView3-build
        593/753 Testing PBGLRandomGraph                  Passed
        594/753 Testing PBGLNamedVertexGraph             Passed
        595/753 Testing TestPBGLGraphSQLReader           Passed
        596/753 Testing TestPBGLGraphSQLReaderFile       Passed
        599/753 Testing TestPBGLCollapseGraph            Passed
        600/753 Testing TestPBGLCollectGraph             Passed
        601/753 Testing TestPBGLPipeline              ***Failed 
        602/753 Testing TestPBGLPedigrees             ***Failed 
        603/753 Testing TestPBGLEdgesPedigrees           Passed
        630/753 Testing TestPBGLAlgorithms               Passed
        
        80% tests passed, 2 tests failed out of 10


I don't know why the example is commented out, perhaps someone at
Kitware would know?

Just in case it's useful, I'm building on an Ubuntu system using
whichever version of OpenMPI that synaptic provides and Boost 1.40.0.

-William


On Thu, 2009-10-01 at 07:40 -0600, Mathieu Malaterre wrote:

> Hi there,
> 
>   I am reading :
> 
> http://www.sandia.gov/Titan/media/Information_Visualization_in_VTK.pdf
> 
>   Looking at page 57, I decided to try the example :
> VTK/Examples/Infovis/Cxx/ParallelBFS.cxx
> 
>   It looks like it was removed from compilation, for a reason I do not
> quite understand:
> 
> http://public.kitware.com/cgi-bin/viewcvs.cgi/Examples/Infovis/Cxx/CMakeLists.txt?r1=1.2&r2=1.3
> 
>   Why can't I just test for the value of VTK_USE_PARALLEL_BGL ?
> 
>   Anyway after applying a small patch I tried it on my computer but
> only got a segfault. Steps:
> 
> $ lamboot
> 
> LAM 7.1.2/MPI 2 C++/ROMIO - Indiana University
> 
> $ ./bin/ParallelBFS
> MPI process rank 0 (n0, p20203) caught a SIGSEGV.
> [2]    20203 segmentation fault  ./bin/ParallelBFS
> 
> Do I need some special command line options ?
> 
> Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20091001/e8a09498/attachment.html>