Overview

IGM runs from the command line. A simulation is fully described by a single YAML configuration file; no source-code editing is needed.

The basic command

igm_run +experiment=params

This tells Hydra to load experiment/params.yaml from the current directory, compose the full configuration, and start the simulation. Run it from the folder that contains the experiment/ directory:

my_experiment/
├── experiment/
│   └── params.yaml
└── [run igm_run here]

If you are using custom modules, place them alongside the experiment folder:

my_experiment/
├── experiment/
│   └── params.yaml
└── user/
    ├── conf/      # Hydra config stubs for each custom module
    └── code/      # Python source files

Tip

Activate your IGM virtual environment before running.

Changing parameters

Any parameter can be changed directly on the command line:

igm_run +experiment=params processes.time.end=2100

To run multiple values in one go (Hydra multirun):

igm_run -m +experiment=params \
  processes.iceflow.physics.sliding.tau_ref=0.05,0.10,0.20

See Hydra: Basics for the full override syntax.

Where to go next

Tutorials

Step-by-step examples

Follow a complete simulation of the Aletsch Glacier from data loading to output — then add Hydra sweeps, particle tracking, and Optuna calibration.

Browse tutorials →

Best practices

Checklist & tips

What to verify before trusting results: numerical stability, domain boundaries, mass conservation, and common pitfalls.

Read the checklist →

Q&A

FAQ

Answers to common questions about installation, configuration, outputs, and troubleshooting.

Browse the FAQ →