What is ParaView?

ParaView (Kitware) is a tool for scientific visualization and post-processing based on VTK. Fields of application are CFD, Astrophysics, Medical Imaging, etc.

License Terms and Usage Conditions

ParaView/VTK is open source, published under a BSD License.

Getting Started

ParaView comprises several tools, which can be used for different setups of systems.

ParaView can be downloaded from ParaView Download page, and is installed on our cluster systems. For opening the interactive GUI on our systems, we recommend the Remote Visualization System at the LRZ.

A more convenient way, using the second option above, is to start pvserver on compute nodes (no GPUs! Software Rendering!), and connect to them via a local GUI through a SSH tunnel.

The fully non-interactive approach via Python scripting and pvbatch is even better adapted to HPC, and can be started MPI parallel, too. The script can be prepared by means of the macro-recording capabilities of the GUI, and then is modified and refined to the needs.

Look for paraview modules on the LRZ Cluster systems via

> module avail paraview

paraview-prebuild are precompiled paraview versions from the paraview.org download page.

Users can download and install any pre-build paraview version desired. It should work out of the box (effort should be minimal). LRZ only provides some selected and tested versions for efficiency reasons. But this means no restriction for users.

Getting Help / Documentation

The ParaView Download page also contains Getting Started and ParaView Guide. We also recommend the ParaVIew Tutorial by Kenneth Moreland (even that for older versions like 5.6 should work) for a starting point.

Users with ambitions should also consult the VTK documentation and programmer's guide.


  1. Jobs (pvserver/pvbatch) crashes: Possibly to few memory available!
    Solution: Increase the number of nodes (possibly decrease MPI tasks per node).
  2. Above problem persists; solution doesn't help: Your data files are not read in parallel, i.e. ParaView does not really exploit parallelism. In the memory monitor that's visible as load imbalance.
    Solution: Use parallel file formats or folder structures (OpenFOAM - decomposed)
  3. ParaView GUI and PV Server cannot be connected: Many reasons possible! Check error message
    1. PV GUI on MAC OS, PV Server on Linux: Problem with Endianess! (Use another OS ... virtualization?; use VNC server on login nodes, and tunnel via SSH to you local client)
    2. Different PV versions of client and server: Download a client of correct version!
    3. Generally: SSH tunnel is maybe not correct; Check for SSH error output