sndust is a collection of python scripts that model the formation of dust grains in a cooling, expanding vapor. Any environment that has time-series temperature and density data can be used as input, but it was orignally designed to study dust production in core-collapse supernovae (CCSNe).

sndust was developed at Los Alamos National Lab (LANL C19146)

The top-level file runtime_settings.json includes several parameters that control the integration of the ODE that models the physics.

• abs_tol, rel_tol control the tolerance of the integration step
• max_dt is the maximum allowed timestep
• *_every are the output options. screen_every is the very much the slowest

The python packages necessary to build are:

• python v3.7
• numpy
• scipy
• numba
• mpi4py
• matplotlib
• h5py
• periodictable

The provided environment.yml file defines an environment that installs these packages using conda

$> conda env create -f environment.yml

sndust can run in either serial or parallel. The runtime configuration must be provided, see runtime_settings.json for an example setup

$> python main.py -c runtime_settings.json

For parallel, invoke using mpi4py.futures

$> mpiexec -n 4 python -m mpi4py.futures main.py -N 4 -c runtime_settings.json