Software

Software available in MPCDF remote visualization sessions

The remote visualization environment has access to all the standard software packages that are available on the cluster in general. In addition, an environment modules section labeled "visualization" is available. It provides a selection of applications and tools for 2D and 3D data visualization. Among those, the widely used workhorses VisIt and ParaView are described briefly below.

Please use the command "module avail" to get an up-to-date list of all the available software, as well as the commands "module show PACKAGE" and "module help PACKAGE" to obtain information aboud a specific package.

 

VisIt

VisIt is a free [and open-source], interactive parallel visualization and graphical analysis tool for viewing scientific data on Unix and PC platforms. Users can quickly generate visualizations from their data, animate them through time, manipulate them, and save the resulting images for presentations. VisIt  can be used to visualize scalar and vector fields defined on two- and three-dimensional (2D and 3D) structured and unstructured meshes. VisIt was designed to handle very large data set sizes in the terascale range and yet can also handle small data sets in the kilobyte range. (Introduction taken from the VisIt website)

  • Website: visit.llnl.gov
  • VisIt documentation page
  • VisIt at the MPCDF
  • INTERACTIVE PARALLEL RENDERING WITH VISIT

    A host profile for the cluster is installed in your VisIt directory (~/.visit/hosts). First, start a VisIt instance:

    $ module load visit
    $ vglrun visit

    To start a parallel engine from the VisIt GUI, select "File -> Open file...". After having chosen the files to be visualized, start a rendering engine. Three options are offered, namely "default", "parallel_threads", "parallel_MPI". Note that parallel rendering may fail if your data format is not adequate.

    NON-INTERACTIVE PARALLEL RENDERING WITH VISIT

    VisIt provides a powerful Python API that can be used to implement complete animations. Typically, the work flow is to first explore data interactively in order to find e.g. useful color scales and view points, and use that knowledge in a second step to develop a VisIt Python script.

    The following code snippet may be used in a VisIt Python script for parallel rendering on a cluster:

    # ---  Python header lines for VisIt parallel rendering
    launchArguments = ("-l", "mpiexec")
    OpenComputeEngine("localhost", launchArguments)
    # --- continue as usual:
    # OpenDatabase(localhost:...,)
    # AddPlot()

    To render in hardware-accelerated mode, use

    $ vglrun visit -cli -noconfig -s yourscript.py.

    To render without any window in software, use

    $ visit -nowin -cli -noconfig -s yourscript.py.

    For further details on VisIt Python scripting, please have a look at the manual on the VisIt webpage.

    TROUBLESHOOTING

    In case you get an error message similar to the following one increase the memory available to VisIt via the submit script and retry.

    VisIt: Message - Rendering window 1...
    VisIt: Error - The compute engine running on drav01 has exited abnormally.
    VisIt: Error - Scalable Render Request Failed (VisItException)
    viewer: Obtained null data reader for rendered image for engine drav01

     

    ParaView

    ParaView is an open-source, multi-platform application designed to visualize data sets of size varying from small to very large. ParaView runs on distributed and shared memory parallel as well as single processor systems and has been successfully tested on Windows, Mac OS X, Linux and various Unix workstations, clusters and supercomputers. (Introduction taken from the ParaView website)

  • Website: www.paraview.org
  • ParaView Wiki
  • Paraview at the MPCDF
  • INTERACTIVE PARALLEL RENDERING WITH PARAVIEW

    For large datasets, you may use several MPI processes to accelerate ParaView. To start the rendering processes, open a terminal window and issue:

    $ module load paraview
    $ mpiexec pvserver

    Then, open a second terminal window and launch the ParaView GUI:

    $ module load paraview
    $ vglrun paraview

    Make sure to connect the GUI to the ParaView server. Please note that parallel rendering does not guarantee increased performance relative to sequential rendering. Speedups highly depend on the dataset.

    Document Actions