FARGO features
Author: Frederic Masset
Here are the main features of the standard public release of the FARGO code.
| Solves Navier-Stokes and continuity equations for a Keplerian disk subject to the gravity of the central object and that of embedded protoplanets |
| Isothermal equation of state, with arbitrary temperature (or sound speed) radial profile |
| 2D polar mesh hydrocode [1] |
| Uses a van Leer upwind algorithm on a staggered mesh. |
| Arbitrary number of planets. [2] |
| Includes the FARGO algorithm (Fast Advection in Rotating Gaseous Objects). [3] |
| Frame is non-inertial and centered on the primary. |
| The full viscous stress tensor of Navier-Stokes equation is implemented. |
| Parallellized both with OpenMP and MPI. [4] |
| Frame is either non-rotating, rotating at a fixed angular velocity, or corotating with a planet or its guiding center [5] |
| A run can be stopped (or killed) and restarted. [6] |
| The disk is non self-gravitating. |
Footnotes
[1] The mesh spacing in azimuth is constant, but the radial spacing can be arbitrary.
[2] The planets are point-like masses. Their trajectory is determined using a fifth order Runger-Kutta integrator.
[3] The code can be run using either the FARGO algorithm or a standard advection technique, for comparison purposes.
[4] FARGO can be run using OpenMP and MPI at the same time, on machines which have nodes of processors with shared memory. Compile time options produce either a sequential or parallel executable from the same source files. Note that the OpenMP version is not maintained, so the use of MPI is recommended. We thank Giovanni Rosotti (LMU, Munich) for providing us with a corrected implementation of the OpenMP version (06/08/2012).
[5] The additional forces that arise in the rotating case are conservatively implemented.
[6] Some parameters can be adjusted before the restart, such as the time sampling of outputs, etc. A parallel run can be restarted as sequential, or with a different number of processors.
RSS 2.0