This year’s Visualize This! challenge focuses on processing and visualizing large datasets with parallel rendering. We recommend using either ParaView or VisIt which were designed with parallel visualization in mind. Both tools are cross-platform, capable of running on a variety of different operating systems (Linux, Windows, and MacOS). Participants are asked to execute their visualization pipeline in a high-performance computing (HPC) cluster environment taking advantage of parallel rendering, and to work either in client-server or batch offscreen rendering mode.

There are other parallel tools you can use besides ParaView and VisIt, and you can even write your own code and decompose the rendering problem with parallel MPI programming. However, these solutions will likely be lower level and hence not recommended for this competition, unless you are visualizing your own data and have a very domain-specific workflow.

Both ParaView and VisIt can be installed as precompiled binaries or compiled from source code. Both offer an interactive GUI and a Python command-line / scripting interface, along with a recording tool to convert GUI actions into Python code. This code can then be modified and converted into a batch processing script to run on a cluster.

If you are not familiar with either ParaView or VisIt, both are quite easy to learn, and we recommend to install and experiment with a local, standalone version on your computer before attempting any parallel rendering. At the bottom of this page we provide several links to documentation and other materials on ParaView and VisIt.

In addition to their visualization capabilities, both ParaView and Visit offer a rich set of exploratory and analysis tools.

ParaView

Started as a collaboration between Los Alamos National Lab and Kitware Inc., ParaView is now primarily developed and maintained by Kitware Inc. ParaView is an open-source, multi-platform data analysis and visualization application capable of reading over a hundred different file formats, including OpenFOAM. To load this competition’s default dataset, you need to open the file case.foam in its top-level directory. This is a dummy, empty file that ParaView uses as a guide to identify and load the OpenFOAM simulation data.

In our Sep-18 webinar Batch visualization on Compute Canada clusters you can see a live demo of interactive client-server visualization of the default dataset on 128 Cedar cores (slides 32-37 and an upcoming video recording) and creating a batch script from this session.

VisIt

Currently developed and maintained by the Lawrence Livermore National Lab, VisIt is an open-source interactive scalable visualization, animation, and analysis tool. VisIt can read over a hundred different file formats too, including OpenFOAM data.

In the latest (3.x) versions of VisIt, we noticed that it can be tricky to make it work with certain Linux distributions, in particular Ubuntu 17 or newer versions. In the same 3.x release, we experienced frequent crashes when drawing plots with remote client-server rendering. For these reasons, we provide only 2.3.x on our clusters. If you are experiencing similar issues, please feel free to contact us.

Other Tools

Blender

Compared to ParaView and VisIt, Blender is more of a 3D artistic visualization and animation tool than a proper scientific visualization package. The quality and style of Blender visualizations can reach Hollywood level, but due to its complexity and a large set of features, the learning curve is quite steep, especially for mastering the more advanced and visually appealing features.

For example, Blender can import 3D scenes from traditional visualization packages and apply photo-realistic rendering. However, unless you have already reached a certain level of proficiency with Blender, we do not recommend it for this competition.

Further Resources: