The AMUN Code

Copyright (C) 2008-2021 Grzegorz Kowal

AMUN is a parallel code to perform numerical simulations in fluid approximation on uniform or non-uniform (adaptive) meshes. The goal in developing this code is to create a solid framework for simulations with support for number of numerical methods which can be selected in an easy way through a parameter file. The following features are already implemented:

• hydrodynamic and magnetohydrodynamic set of equations (HD and MHD),
• both classical and special relativity cases for the above equations,
• Cartesian coordinate system,
• uniform and adaptive mesh generation and update,
• 2nd to 4th order time integration using Strong Stability Preserving Runge-Kutta methods,
• 2nd order TVD interpolation with number of limiters and higher order reconstructions,
• Riemann solvers of Roe- and HLL-types (HLL, HLLC, and HLLD),
• standard boundary conditions: periodic, open, reflective, hydrostatic, etc.
• turbulence driving using Alvelius or Ornstein–Uhlenbeck methods,
• viscous and resistive source terms,
• support for passive scalars (up to 100),
• data stored in internal XML+binary or HDF5 format,
• support for Zstandard, LZ4, and LZMA compression in XML+binary format,
• Python interface to read snapshots in both formats,
• MPI parallelization,
• completely written in Fortran 2008,
• simple Makefile or CMake for executable building,
• minimum requirements, only Fortran compiler and Python are required to prepare, run, and analyze your simulations.

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.

Developers

• Grzegorz Kowal grzegorz@amuncode.org

Requirements

• Fortran 2003 compiler, tested compilers include:
  • GNU Fortran version 4.5 or newer,
  • PGI Community Edition, version 18.10 or newer,
  • Intel Fortran compiler version 9.0 or newer.
  • NVIDIA HPC compiler version 21.9. Warning: I could not make it run with the included MPI libraries.
• Optional, but recommended, OpenMPI for parallel runs, tested with version 1.8 or newer.
• Optional support for XML-binary format compression requires: LZ4 library, Zstandard library, or LZMA library
• Optional HDF5 libraries, tested with version 1.10 or newer. The code now uses the new XML-binary snapshot format. However, if you still want to use older HDF5 snapshot format, you will need these libraries.
• Optional CMake version 3.16 or newer, for managing the build process.

Environment Variables

If you need to use the HDF5 libraries and they are not installed in the default location, i.e. in the system directory /usr, make sure that the environment variable HDF5DIR is set in your ~/.bashrc (or ~/.cshrc) and pointing to the location where the HDF5 libraries have been installed.

Recommended compilation (using CMake)

1. Clone the AMUN source code:

• from Bitbucket: git clone https://grzegorz_kowal@bitbucket.org/amunteam/amun-code.git,
• from GitLab: git clone https://gitlab.com/gkowal/amun-code.git
• or unpack the archive downloaded from page Downloads.

2. Create a directory for compilation in any location, e.g. mkdir cmake-build && cmake-build.
3. Call ccmake <path to amun-code>, e.g. ccmake .., and press 'c' once. Configure available options. Press 'c' once again, and 'g' to generate makefiles. Alternatively, just call ccmake <path to amun-code> for default options.
4. Compile the code using make. The executable file amun.x should be created.

Alternative compilation (using make.conf)

1. Clone the AMUN source code:

• from Bitbucket: git clone https://grzegorz_kowal@bitbucket.org/amunteam/amun-code.git,
• from GitLab: git clone https://gitlab.com/gkowal/amun-code.git
• or unpack the archive downloaded from page Downloads.

2. Go to directory build/hosts/ and copy file default to a new file named exactly as your host name, i.e. cp default $HOSTNAME.
3. Customize your compiler and compilation options in your new host file.
4. Go up to directory build/ and copy file make.default to make.config.
5. Customize compilation time options in make.config.
6. Compile sources by typing make in directory build/. The executable file amun.x should be created there.

Usage

In order to run some test problems you can simply copy the problem parameter file from directory problems/ to the location where you wish to run your test. Copy the executable file amun.x from the build/ directory compiled earlier. If you provide option -i <parameter_file>, the code will know that parameters have to be read from file <parameter_file>. If you don't provide this option, the code assumes that the parameters are stored in file params.in in the same director.

In order to run serial version, just type in your terminal: ./amun.x -i ./params.in.

In order to run parallel version (after compiling the code with MPI support), type in your terminal: mpirun -n N ./amun.x -i ./params.in, where N is the number of processors to use.

Reading data

By default, the code uses new XML+binary snapshot data format. It can also be forced by setting parameter snapshot_format to xml.

In order to read produced data in this format, you will need to install the provided Python module. Simply change to python/ directory and run python setup.py install --user to install the module in your home directory.

Import the module in your python script using from amunpy import *, and then initiate the interface using snapshot = AmunXML(<path to the snapshot directory>) and read desired variable using var = snapshot.dataset(<variable>).

The function dataset() returns rescaled uniform mesh variable as NumPy array.

If you want to read data from HDF5 snapshot, just use var = amun_dataset(<snapshot HDF5 file>, <variable>).