[Paraview] How to create parallel PVTU files?

[Christoph Moder]

Hi!

We would like to process large datasets on a cluster. Our software 
writes only monolithic *.vtk files, and I think that makes Paraview slow 
because large amounts of data have to be transferred over the network. 
So I think it would be a good idea to divide the data files into 
subgrids and store them in the PVTU format. (Correct me if I'm wrong.) 
But how can I do that? I have read about the vtkDistributedDataFilter; 
can I load the data, apply the DistributedDataFilter, and write the 
subgrids as VTU files, connected by a PVTU file? Any hints?

Thanks,
Christoph