Instructor Notes

This is a placeholder file. Please add content here.

Introduction


Instructor Note

The primary goal of this session is to present potential use cases for JUBE and enable participants to apply these insights to their own workflows on HPC systems.



Instructor Note

It is important for the learners to understand that JUBE itself does know what makes your workflow reproducible on its own. To create a reproducible workflow learners will have to identify which information and actions make the specific workflow reproducible, and use JUBE to automatically collect and execute those, respectively.



An example application


Instructor Note

We use GROMACS for this example, as it is a known code in the HPC community, it is readily available, and from installation to execution needs several steps in a workflow.



Working with JUBE


An initial workflow specification


Influencing parameters with tags


Instructor Note

You can demonstrate the workflow run and showcase that the new archive was downloaded, and the software eventually installed in a different directory.



Spanning parameter spaces


Using templates


Instructor Note

This episode introduces the simple text-based substitution system of JUBE to create text-based input files or batch scripts for submission to an HPC workload manager.



Instructor Note

We end this without an exercise, as the next episode will include the generation of batch scripts using the above steps.



Including external data


Running an application


Output analysis and presentation


Thoughts on reproducibility


Wrap-up and outlook