PDS training¶

Introduction¶

In this PDS training we introduce you to the basic concepts and features of the PDS ecosystem. The training is intended to be worked through in order, as more complicated topics will be covered in the later sections.

Important

This training assumes you have access to ITER's SDCC, as all dependencies will be available as modules. Moreover, you will need access to a graphical environment to visualize the simulation results. It is recommended to follow this training through the NoMachine client.

After this, the separate Advanced PDS training track is aimed at developers: it shows how to build your own workflow out of existing support and simulation actors, and how to write a new actor of your own. The rest of this training is relevant to any PDS user.

We assume you have some basic knowledge of IMAS. You will also need some basic familiarity with Python. For a refresher, see the Python tutorial. We also assume some basic knowledge of MUSCLE3. For more information on the visualization actor, see the IMAS-MUSCLE3 training. For more information on the waveform-editor, see the Waveform-Editor training.

Training contents¶

The sections are meant to be followed in order:

Setting up the PDS: Getting the PDS and the scenario data ready on SDCC
Understanding the PDS: Learn what workflows, scenarios, cases and runs are, why the PDS keeps them apart, and how to find out what is available
Running your first PDS workflow: Run a small workflow, follow it live in the MUSCLE3 dashboard, work out what went wrong when one fails, and submit a run to a compute node
Running existing workflows: Run the premade ITER cases, watch them while they go, and read the results they leave behind
Configuring existing workflows: Change the pulse design, the solver settings and the time window of a case

See Advanced PDS training for a separate, follow-on training aimed at developers: it shows how to assemble a workflow out of existing actors step by step, profile a run, create a new actor of your own, bring a new scenario into the PDS, and visualize the data flowing between actors while a workflow runs.