Skip to content
/ EcoSIM Public
forked from jinyun1tang/EcoSIM

EcoSIM land/surface process model - fork to improve CMake build system

Notifications You must be signed in to change notification settings

amanzi/EcoSIM

 
 

Repository files navigation

EcoSIM

A biogeochemical modeling library spins off the ecosys model.

Download

git clone --recursive git@github.com:jinyun1tang/EcoSIM.git

Building

Use the script in the root directory to build the code:

sh build_EcoSIM.sh

or run it as a bash script:

./build_EcoSIM.sh

This should do everything including building and linking all the required tpls, and configuring, building and linking EcoSIM itself.

The executable will be found in ./build/bin/ecosim.f90.x

This has been tested with gcc on multiple systems, but not intel compilers. See the .github/workflow/ecosim-ci.yml file for examples. You will need cmake to use the build_EcoSIM.sh script.

Additional options

The build_EcoSIM.sh script lets you set some opttional parameters:

debug=0 mpi=0 shared=0 verbose=0 sanitize=0 regression_test=0

precision="double" prefix="" systype=""

#Leave empty to just use the environment variable compiler CC="" CXX="" FC=""

You can also set options via the command line, for example:

./build_EcoSIM.sh CC=/path/to/cc CXX=/path/to/cxx FC=/path/to/fortran --debug --regression_test

Run

./build_EcoSIM.sh --help

for a full list of optional arguments

using docker on mac

  1. make sure docker is installed on your machine, please see https://www.docker.com, and ask for help from whatever chatbot you have access to.

  2. builidng the container

docker build -f docker/ubuntu-compiler.dockerfile -t my-compiler .

  1. run the container

docker run -it -v $(pwd):/EcoSIM -w /EcoSIM my-compiler

this will allow you to access everything under the EcoSIM directory in the container

About

EcoSIM land/surface process model - fork to improve CMake build system

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Fortran 79.8%
  • Jupyter Notebook 14.7%
  • Python 4.1%
  • CMake 1.3%
  • Shell 0.1%
  • Gnuplot 0.0%