Simulations with AREPO

AREPO is a moving-mesh hydrodynamics code designed for cosmological and astrophysical simulations. It combines the adaptive resolution of smoothed-particle hydrodynamics (SPH) with the accuracy of grid-based methods, providing excellent performance for both cosmological and galaxy-scale problems.

Pisces includes native support for generating AREPO-compatible initial conditions through the arepo extension module. The resulting output is an HDF5 file formatted according to AREPO’s IC conventions and represented within Pisces by the AREPOParticleDataset.

Detailed installation instructions for AREPO are provided in the official AREPO User Guide. For Pisces compatibility, several compile-time and runtime options must be adjusted in your AREPO build configuration. See Compile-Time Settings for a summary.

Overview of Pisces Support

The arepo module provides all necessary tools for creating, managing, and validating AREPO initial conditions. The main entry point is the AREPOFrontend, which wraps an existing Pisces initial condition file and produces AREPO-readable particle datasets.

🧭 Workflow Steps

1. Build models — Using Pisces, generate whatever models you’d like to use in the simulation. This might be galaxy mergers, cosmological models, etc. See Astrophysics Models in Pisces for an overview of Pisces model generation. These are later composed into a set of initial conditions for the simulation.

2. Assemble initial conditions — Once the models are generated, they can be combined and oriented in space to produce a set of initial conditions (InitialConditions). For a detailed description of the different types of initial conditions available and how they work, see the documentation on initial conditions: Initial Conditions for Simulations.

3. Generate particles — Once you have constructed your initial conditions, it will be necessary to to convert the constituent models into particle datasets. This can be done through the initial conditions class you generated in the previous step; however, not all models can generate particles. See the examples in Pisces Example Gallery for examples.

4. Initialize the frontend — Wrap the resulting initial conditions with AREPOFrontend.

5. Write AREPO ICs — Call generate_initial_conditions() to produce HDF5 files compatible with AREPO’s input routines.

⚠️ Compatibility Reminder

AREPO’s compile-time and runtime flags must exactly match those used by Pisces when generating initial conditions.

• If AREPO is compiled with 32-bit particle IDs, Pisces must also use 32-bit IDs (long_ids=False).

• If AREPO is compiled with MHD or cooling enabled, the corresponding fields (MagneticField, ElectronAbundance) must be present in the ICs.

• Dimensionality (1D, 2D, 3D) must match exactly between Pisces and AREPO.

The AREPO Frontend

Like all simulation codes supported by Pisces, AREPO is accessed through a frontend. The AREPOFrontend handles creation, validation, and configuration of AREPO-compatible initial condition files.

To create a frontend:

from pisces.extensions.simulation.arepo.frontends import AREPOFrontend
from pisces.extensions.simulation import InitialConditions

ic = InitialConditions("path/to/initial_conditions_file")
frontend = AREPOFrontend(ic)

This command generates a configuration file named AREPOFrontend_config.yaml in the same directory as your initial conditions. You can edit it manually or programmatically via the frontend’s configuration API.

To inspect or modify settings in Python:

>>> print(dict(frontend.config["makefile"]))
{'simulation_type': '3D', 'number_of_particle_types': 6, 'nbits_id': 64, 'double_precision': True}

To locate the YAML configuration file on disk:

>>> print(frontend.config_path)
PosixPath(".../path/to/AREPOFrontend_config.yaml")

Compile-Time Settings

Certain AREPO Makefile options determine the binary layout and precision of the simulation files. These are mirrored in the frontend configuration under the makefile.* section.

Relevant Compile-Time Settings

Pisces Setting	Makefile Flag	Description
makefile.simulation_type	TWODIM / ONEDIM / ONEDIMS_SPHERICAL	Sets the simulation dimensionality. Must match the IC file.
makefile.number_of_particle_types	NTYPES / NTYPES_ICS	Number of particle types. Must be ≥ number in ICs.
makefile.nbits_id	IDS_32BIT / IDS_64BIT	Particle ID bit depth. Must match the IC file.
makefile.double_precision	DOUBLEPRECISION / INPUT_IN_DOUBLEPRECISION / READ_COORDINATES_IN_DOUBLE	Internal runtime precision for floating-point arithmetic. This setting determines the precision of the initial conditions. We do not support mismatched precision between coordinates and other fields.
physics.cooling	COOLING	Enables radiative cooling. Requires ElectronAbundance field in ICs.
physics.mhd	MHD	Enables magneto-hydrodynamics. Requires MagneticField field in ICs.
physics.passive_scalars	PASSIVE_SCALARS=<N>	Number of advected passive scalars to track.

Runtime Parameters

AREPO’s runtime parameters specify the physical and cosmological context of the simulation. These correspond to the parameters.* section of the frontend configuration.

Relevant Runtime Settings

Pisces Setting	AREPO Parameter	Description
parameters.boxsize	BoxSize	Simulation box size in code units.
parameters.units.length, mass, velocity	UnitLength_in_cm, UnitMass_in_g, UnitVelocity_in_cm_per_s	Physical unit definitions (time is derived from L/V).
parameters.cooling / parameters.mhd	CoolingOn / MHDOn	Enable or disable cooling and MHD at runtime.
parameters.ic_format	ICFormat	Must be set to 3 for HDF5 ICs.

Field Configuration

Finally, the frontend defines how Pisces particle fields are mapped to AREPO fields. This mapping is provided in the model_template section of the configuration file.

Example model_template

model_template:
  ParticleType0:
    name: gas
    fields:
      Coordinates: particle_position
      Velocities: particle_velocity
      ParticleIDs: particle_id
      Masses: particle_mass
      InternalEnergy: particle_internal_energy
  ParticleType1:
    name: dark_matter
    fields:
      Coordinates: particle_position
      Velocities: particle_velocity
      ParticleIDs: particle_id
      Masses: particle_mass
  ParticleType4:
    name: stars
    fields:
      Coordinates: particle_position
      Velocities: particle_velocity
      ParticleIDs: particle_id
      Masses: particle_mass

Optional fields are added automatically based on enabled physics: - MagneticField (for MHD) - ElectronAbundance (for cooling) - PassiveScalars (if passive_scalars > 0)