Sailfish

Sailfish is a GPU-accelerated astrophysical gasdynamics code.

The main repository is hosted here: github.com/clemson-cal.

Quick-start

Basic use from the command line is like this: bin/sailfish shocktube --end-time=0.2 --resolution=1000. This command will run a built-in setup called shocktube to a simulation time of 0.2 seconds, on a 1d grid with 1000 zones. The command line tool will print messages to the terminal, and write a file called chkpt.final.pk to the disk in the current working directory when the simulation finishes. This file is a Python pickle, containing the simulation state and hydrodynamics solution data, and you can load it in Python to plot the data.

Sailfish can also be used as a Python module from a custom script, as in the code below:

from sailfish.driver import run
from matplotlib import pyplot as plt

state = run("shocktube", end_time=0.2, resolution=1000)
rho = state.solver.primitive[:, 0]
plt.plot(rho)
plt.show()

The run function takes a sequence of arguments to control the driver and problem setup, and then returns the simulation final state. However no side-effects are performed by run and the terminal output is suppressed (you can re-enable it with the keyword argument quiet=False).

API documentation

`sailfish.driver`	Library functions and command-line access to the simulation driver.
`sailfish.event`
`sailfish.kernel`	A Python module to facilitate JIT-compiled CPU-GPU agnostic compute kernels.
`sailfish.mesh`	Contains classes for different mesh geometries and dimensionalities.
`sailfish.physics`	Physics support modules, other than hydro solvers.
`sailfish.quad_tree`
`sailfish.setup_base`	The Setup class dictates physics and solver choices to the driver.
`sailfish.setups`
`sailfish.solver_base`	The Solver class maintains physics state and implements a solution scheme.
`sailfish.solvers`	Physics solvers and solution schemes.
`sailfish.subdivide`	Utility functions for domain decomposition.