This half-day seminar introduces five workflow management tools for automating computational research on Perlmutter: GNU Parallel, signac, Maestro, Merlin, and AiiDA.
Participants will run hands-on examples on Perlmutter that demonstrate how each tool addresses a distinct challenge in managing large-scale scientific computations.
By the end of the seminar, participants will be able to identify which tool fits their workflow requirements and apply it to their own research.
Topics
The seminar will cover the following tools.
- GNU Parallel: Execute independent tasks in parallel across cores or nodes without writing job scripts.
- signac: Organize parameter spaces and track simulation inputs and outputs with a lightweight Python framework.
- Maestro: Define multi-step workflows as directed acyclic graphs with explicit task dependencies.
- Merlin: Coordinate large-scale ensemble simulations across distributed resources using a producer-consumer model.
- AiiDA: Automate and track computational workflows with full provenance, enabling reproducible research.
Prerequisites
To get the most out of this seminar, participants should be familiar with
- Submitting jobs on a Slurm-based HPC system,
- Shell scripting (Bash), and
- Basic Python (reading and running scripts).
How to attend
Registration for this event is open to all.
If you are not a current NERSC user, you must register by May 14, 2026, to give us time to set up your temporary training account on Perlmutter.
Current NERSC users should register before 12 p.m. PDT on May 27, 2026.