[Paraview] need/share parallel 32-partitions data

[Berk Geveci]

> There seems to be two approaches to parallel visualation.
> 1-The paraview file contains the global mesh and say 32 processors are used
> to visualize the global data. Maybe I am confused but this would seem to be
> the approach that you suggest above. Else, how would I split the above into
> partitions?
> 2-The approach we use is we want to avoid writing global mesh files and a
> global .vtu file and instead write a .vtu file for each processor. For a
> computation using 256 processors, 256 separate vtu files are written at each
> time step with each file being approximately 1/256 the size of a global
> file. The idea is to use say 32 paraview processors to visualize the 256 vtu
> files at each time step.

Perfect. This is the right way of doing it. It should work with any
number of nodes and partitions as long as you have enough memory. Is
there a dataset of small to medium size that you can share so that I
can track down the problem? It would be best if I can open it on at
most 8 processors...

-berk