AMA27's training is organised in three tracks. A module is a short course. Each is taught as a workshop or a webinar series, and written so you can work through it alone. The tutorials listed under a module are single lessons, 5 to 15 minutes each, that you can take whenever you need them.
Modules marked in draft have an outline but no dates yet. Lessons without a link are still being written.
A. Onboarding
- Basics of using AMA27 (in draft): your account, the shell, where files live, and a first job.
- Using Slurm (in draft): asking for cores, memory and time; finding out why a job waits or fails; job arrays.
- Open OnDemand (in draft): using the cluster from a browser, for files and for interactive apps.
B. Productivity & Data
- Advanced shell (in draft): scripting, text processing, and running many small tasks.
- Data management (in draft): where data lives, what to keep, and how to describe it.
- Data transfer (in draft): moving large data on and off the cluster.
- HDF5 (in draft): a portable, self-describing format for large scientific data.
- Visualization: seeing simulation output, molecules and volumes, and making the same picture again by script.
- Start ParaView on OnDemand
- Draw streamlines in ParaView
- See orbitals and electron density in ParaView
- Comparisons of the programs that can do one lesson: streamlines, orbitals and electron density
C. Parallel & Performance
- OpenMP (in draft): threads on one node.
- MPI (in draft): processes across nodes.
- Profiling & optimization (in draft): measuring where a program spends its time, and fixing what you find.
Where to start
Do Start an OnDemand session once, then take the module that matches your work. Do the A track first. It assumes no cluster experience, while B and C assume you can submit a job.