Containerized Visualization Tools
Visualization of a hemodynamics simulation. The data is partitioned in four MPI ranks and colored by rank ID. Rendered on Argonne Leadership Computing Facility resources at Argonne National Laboratory. Data courtesy Prof. Amanda Randles (Duke University) and team.
Docker and Singularity are well known containerization platforms, becoming increasingly popular in high performance computing. They allow for complex installation processes to be performed once; the resulting environment can be executed later on other systems. Scientific visualization tools, such as ParaView, depend on a significant number of libraries that may be challenging to build and deploy on new systems. Moving ParaView and its libraries into a container would simplify the deployment of the installation, handling dependencies independently of the system in use. Through this work, we explore the process for containerizing ParaView and evaluate its performance on a number of distributed parallel computers against their native installations.
The ddiLab Container Team:
This research used resources of the Argonne Leadership Computing Facility, which is a U.S. Department of Energy Office of Science User Facility supported under Contract DE-AC02- 06CH11357.