mikbitz/MadHPC
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
This distribution contains a version of the Madingley global ecosystem model derived from the original C# code implemented by:- Emergent Global Patterns of Ecosystem Structure and Function from a Mechanistic General Ecosystem Model , Michael B. J. Harfoot , Tim Newbold, Derek P. Tittensor, Stephen Emmott, Jon Hutton, Vassily Lyutsarev, Matthew J. Smith, Jörn P. W. Scharlemann, Drew W. Purves PLOS Biology 12(4) 2014 e1001841 https://doi.org/10.1371/journal.pbio.1001841 The code here has been adapted to run in parallel either on a standalone machine or across a cluster using the MPI libraries and a modified version of repastHPC v2.2.0 downloaded from https://repast.github.io/repast_hpc.html Build instructions can be found in INSTALL.txt Once you have the mad_model executable, depending on how your system is set up you may then need to set the system variable LD_LIBRARY_PATH in order for the model to find shared libraries - see e.g. the file LIBS which also makes sure the path to the MPICH executalbes is set. You can check that all shared libraries can be found with:- ldd ./mad_model Run the model with the following incantation:- mpi_run -np <no of threads> ./mad_model config.props model.props The value in <no of threads> MUST be consistent with the total number of threads specified in model.props or the run will immediately crash. e.g. if proc.per.x = 3 proc.per.y = 2 then use mpi_run -np 6 ./mad_model config.props model.props Parameters in model.props can also be changed on the command line e.g. mpirun -np 1 ./mad_model config.props model.props proc.per.x=1 proc.per.y=1 stop.at=2 See model.props for a list of model parameters. For the behaviour of config.props consult the repast_hpc documentation.
About
MPI version of the madingley model based on RepastHPC v2.2.0
Resources
Stars
Watchers
Forks
Packages 0
No packages published