Testing scientific software on HPC often times involves dozens or hundreds of simulation runs. Therefore it is important to have all the data well organized and to be able to quickly evaluate data.
Ideally you can “live” watch your results on your HPC. This enables live inspection of the parameter studies and stopping/re-running those not creating satisfactory results to save computational resources.
This picture shows how parameter studies can be arranged and viewed via the browser on a HPC. The process is well compatible with cross-linking and containerization, as the secondary data is stored directly with the respective simulation and can be used in a publication.
If you want to read more about Jupyter, please refer to our literature.
To start working you first need a CI-pipeline and tests.
You can also get a list of our articles assigned to this chapter via the tags on the top of every page.