[Paraview] Parallel XML VTK with arbitrary polyhedra

[Le Chenadec, Vincent]

Hi,

I figured using a PVD file does the job, for example

<VTKFile type="Collection" version="1.0" byte_order="LittleEndian" header_type="UInt64">
  <Collection>
    <DataSet part="0" file="sphere_0.vtu"/>
    <DataSet part="1" file="sphere_1.vtu"/>
    <DataSet part="2" file="sphere_2.vtu"/>
    <DataSet part="3" file="sphere_3.vtu"/>
    <DataSet part="4" file="sphere_4.vtu"/>
    <DataSet part="5" file="sphere_5.vtu"/>
    <DataSet part="6" file="sphere_6.vtu"/>
    <DataSet part="7" file="sphere_7.vtu"/>
  </Collection>
</VTKFile>

I’m still interested in getting .pvtu files to work though, if anyone has an idea of how to get it to work with polyhedral cells…

Best,
Vincent

From: "<Le Chenadec>", Vincent Le Chenadec <vlc at illinois.edu<mailto:vlc at illinois.edu>>
Date: Monday, January 5, 2015 at 2:40 PM
To: "paraview at paraview.org<mailto:paraview at paraview.org>" <paraview at paraview.org<mailto:paraview at paraview.org>>
Subject: Parallel XML VTK with arbitrary polyhedra

Hi there,

I’m attempting to write a parallel XML file to source unstructured data to read in Paraview (written from an in-house fortran CFD code). Everything works out of the box when I use zoo elements (hexahedra, tetrahedra, …), not so when using arbitrary polyhedra.

I’m attaching a sample test (a sphere discretized on a 16x16x16 lattice, see sphere1.png). The .vtu files I dump seem to be valid, for the external boundaries are displayed accurately when visualizing them separately (see attached sphere2.png to see the partitioning).

When loading the .pvtu file, however, the external boundary considers only the zoo elements (hex only, in this case), not the polyhedral cells. Paraview also crashes whenever I apply a filter, which suggests I’m doing something wrong.

I’m most likely missing something in the .pvtu file? I tried a few things, e.g. adding a <PCells/> element where with <PDataArray/>’s referencing the faces and faceoffsets arrays specific to polyhedral cells, but no success so far… Any input is much welcome at this point!

Thanks,
Vincent

PS: I’m using the 4.2.0 built of Paraview. The ASCII .vtu and .pvtu files I use are attached in a tarball.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20150105/c0ebd0b5/attachment.html>