mugy

Multiscale gyrofluid code

Written in C and CUDA, with postprocessing scripts in Python.

Installation

The code has the following dependencies:

• GSL (GNU Scientific Library)
• CUDA
• MPI
• FFTW
• ADIOS2

Follow these steps to build mugy:

1. Specify paths to include and library directories, as well as some flags, for each of these dependencies in the setup.sh script.
2. Run the setup script: . ./setup.sh.
3. Run the makefile: make.

This produces the executable mugy.x.

Runing mugy

mugy must be run with MPI and requires at least 2 command line inputs, the full path to the input file and the full path to the output directory. For example, to use the input file mugy_test.in and output to /scratch/jdoe/test/ use:

mpirun -np 1 ./mugy.x ./mugy_test.in /scratch/jdoe/test/

An optional third command line argument may be specified: the restart directory (the directory of a previous simulation that we wish to continue). For example, if the restart directory is /scratch/jdoe/start/ then this restart simulation may be launched with

mpirun -np 1 ./mugy.x ./mugy_test.in /scratch/jdoe/test/ /scratch/jdoe/start/

Developer notes

Please try to use following naming conventions whenever possible:

• Datatype names must end with _t.
• enum options must be capitalized.
• append mugy_ as a prefix to structs, datatypes, and functions that may be publicly visible.