Interactive Visualization at DKRZ

Using NCL and Markers to plot data on a grid Overlaying a regional on a global grid Multi-Variable 3D plot with ParaView


Postprocessing and 2D visualizations can be done on a partition of Mistral reserved for data processing and analysis. For interactive 2D plotting in Python, R, and Julia, we provide a Jupyterhub server. Additionally to that, mistral also comprises a visualization partition equipped with powerful GPUs. The GPU nodes here are especially designed for interactive 3D visualization and analysis of large model data. We provide a wide range of Visualization Software on mistral.

Visualization Software

On Mistral, all our visualization software is made available using modules. The command module avail provides you with a list of all software modules that are available.

2D Data Analysis and Visualization

For other software that is installed on mistral such as further 2D visualization solutions please refer to our software list. Additionally to that, several 3D visualization solutions which require GPUs to run are available:

3D Visualization and Data Analysis

Image and video processing

Access and reservation of a GPU node

Our supercomputer Mistral currently includes 21 GPU nodes. This built-in visualization server is described in more technical detail here. These nodes are equipped with the same CPUs and a very similar software tree as the supercomputer. On these nodes, you can start a VNC server and export a full desktop to your personal system. Please follow the instructions in Accessing Mistral’s GPU Nodes to log in to them and start the VNC software. See Remote 3D Rendering for more details on the TurboVNC/VirtualGL based remote rendering technique, which enables using Mistral’s GPU nodes from virtually anywhere.

List of blog posts