On Thu, Dec 11, 2008 at 1:39 PM, Jacques Papper <[email protected]> wrote: > I have a 100 million point connected graph (No cell information - only > points and edges and boundaries defined by certain points and edges) that I > would like to visualize. I would like to be able to load it in parallel and > post-process it in parallel. I would like to do things like streamlines / > surface contours on the boundaries of the graph etc... Also I would like it > to have groups (or blocks) for the different surfaces in the graph. Is there > any way of doing this with ParaView ? Is there a certain class / algorithm > that can accomodate this type of data structure ?
Sandia National Labs and Kitware are in the midst of making it possible to store and process very large graphs in distributed memory using vtkGraph and the Parallel Boost Graph Library. Some of this is currently in VTK, but it is not yet incorporated into ParaView. Your other choice which may work for you is to use vtkPolyData with line cells, since some of the operations you are talking about exist in ParaView for polygonal geometry, although I don't know the exact sequence of filters you would need for your use case. ParaView also has the notion of multi-block data sets which can group assoicated data sets into one object. Jeff _______________________________________________ ParaView mailing list [email protected] http://www.paraview.org/mailman/listinfo/paraview
