Dupost

Dupost is an implementation of a coupled DUal POrosity STokes (dupost) model.

Publications:

1. Douglas, C. C., Hu, X., Bai, B., He, X., Wei, M., & Hou, J. (2018). A Data Assimilation Enabled Model for Coupling Dual Porosity Flow with Free Flow. In Proceedings - 2018 17th International Symposium on Distributed Computing and Applications for Business Engineering and Science, DCABES 2018. https://doi.org/10.1109/DCABES.2018.00085

2. Hu, X., & Douglas, C. C. (2019). An Implementation of a Coupled Dual-Porosity-Stokes Model with FEniCS. In J. M. F. Rodrigues, P. J. S. Cardoso, J. Monteiro, R. Lam, V. V Krzhizhanovskaya, M. H. Lees, … P. M. A. Sloot (Eds.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11539 LNCS, pp. 60–73). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-22747-0_5

Installation

The implementation depend only on FEniCS. Here we introduce two ways of installing FEniCS, which should cover most cases. For a complete list of methods, including building from source, see FEniCS Download Page.

Docker (recommended)

For Unix (Linux/MacOS) Users

1. Install Docker Desktop.

2. Run the following command in your favorite terminal

shell curl -s https://get.fenicsproject.org | bash

1. Download and extract (or clone) dupost. cd into the result folder, then run

shell fenicsproject run

This will bring you into a docker container with fenics installed, as you can tell from the shell prompt, which looks like fenics@c058286e25c5:~/shared$. Only your working directory before entering the container can be accessed inside the container in ~/shared folder. Read FEniCS in Docker and Docker docs for more information.

1. Inside the docker image, run

shell pip install -e . --user

which will install dupost. You can check your installation by running

shell pytest-3 ./test

For Windows Users

1. Install Docker Desktop.

Spack

This method does not requrie a root access.

1. Install Spack by following the installation guide. Add spack/bin to your PATH by following the instruction or use the full path to spack/bin/spack for any following spack command.

2. Configure compilers with the command

shell spack compiler find

1. Install FEniCS using Spack:

shell spack install fenics@2019.1.0.post0

This will take a long time (hours on an early 2015 MacBook Pro). See Spack Documentations if you want to customize the build. On some supercomputers, it is necessary to use a pre-installed MPI. This is configured in the file ~/.spack/packages.yaml. If you need to use an OpenMPI of version 1.10.1 installed in /opt/local for example, the packages.yaml should contain the following lines

yaml packages: openmpi: paths: openmpi@1.10.1: /opt/local buildable: False

And you should consider using the pre-installed compiler used to build that MPI for building FEniCS. See System Packages for detail.