Installing Prerequisites on Ubuntu

There are a few prerequisites which must be installed on your machine before you will be able to build and install OpenMOC. All of the prerequisites can easily be installed using a standard package manager, such as apt-get for Ubuntu. The following instructions detail which packages are needed, with instructions on how to install them using apt-get.

The following command will install all required and optional dependencies on Ubuntu 12.04 or later:

sudo apt-get install build-essential git swig python-dev python-numpy python-matplotlib python-h5py python-pillow

A preferred alternative to apt-get is using anaconda to install Python packages. Anaconda will automatically manage the package versions, ensuring compatibility. Anaconda will also allow you to install packages on a system where you don’t have admin priviledges, like most computing clusters. Once you have Anaconda (or Miniconda installed), you can install all OpenMOC’s Python dependencies with:

conda config --append channels conda-forge
conda install swig numpy matplotlib h5py pillow

Required

  • GNU’s C++ compiler

    In order to compile OpenMOC, you will need to have a C++ compiler installed on your machine. It is recommended that you build OpenMOC with g++ version 4.4 or later. To install g++ on Debian Linux or a Debian derivative such as Ubuntu, use the following command in the console:

    sudo apt-get install build-essential

  • Python

    OpenMOC extensively uses Python for rapid data processing and visualization. OpenMOC uses a mixture of fast, compiled C/C++/CUDA code with Python bindings such that users are entirely insulated from having to write in C/C++ or CUDA, and can simply write a Python script with calls to OpenMOC.

    Currently, OpenMOC has been tested with Python versions 2.6, 2.7, 3.1 to 3.7. Even if you already have Python installed, it is recommended that you install it again using apt-get to ensure that it is properly configured with g++. You can easily install version 3.7 in Ubuntu as follows:

    sudo apt-get install python3.7

    In addition, you must install the Python development headers:

    sudo apt-get install python-dev

  • Git

    The OpenMOC development team uses the Git version control software for code development and collaboration. To download the code from our GitHub repository, you must first install git. You can do so on Ubuntu as follows:

    sudo apt-get install git

  • SWIG

    The Simplified Wrapper Interface Generator (SWIG) is used to wrap the compiled C/C++ and CUDA source code and generate Python bindings. SWIG can be installed on Ubuntu as follows:

    sudo apt-get install swig

  • NumPy

    NumPy is a commonly used, general purpose numerical array creation and manipulation package for Python. OpenMOC uses NumPy to seamlessly transfer numerical data between Python and the C/C++ back-end compiled code. To install NumPy on Ubuntu, use the following command:

    sudo apt-get install python-numpy

Optional

  • matplotlib

    The matplotlib Python package is a general purpose utility for plotting and visualizations. OpenMOC uses matplotlib in the openmoc.plotter module to generate plots, such as flux and power distributions. Although matplotlib and the openmoc.plotter modules are not required for an OpenMOC simulation, visualizations can be a helpful tool in debugging and input model validation. To install matplotlib on Ubuntu, issue the following command:

    sudo apt-get install python-matplotlib

  • h5py

    The h5py Python package contains tools to create, retrieve, and manipulate data stored using the HDF5 binary format. OpenMOC leverages the h5py functionality in its openmoc.materialize module which includes routines to import/export multi-group nuclear cross-section data to/from HDF5. In addition, the openmoc.process module contains routines to import/export simulation and performance data to/from HDF5 output files. Although h5py and the openmoc.materialize and openmoc.process modules are not required to run OpenMOC, they can facilitate fast access and storage of binary data.

    To install h5py on Ubuntu, issue the following command:

sudo apt-get install python-h5py

  • pillow

    The PIL (or its fork pillow) package is used to compare images in the test suite, to check that the plotter module is working properly. It is not required outside of the test suite. To install PIL or pillow on Ubuntu, issue the following command:

sudo apt-get install python-pil

  • scipy

    The scipy package is used for the Krylov solver. It is not required for running a regular MOC (+CMFD) solve To install pillow on Ubuntu, issue the following command:

sudo apt-get install python-scipy

  • mpi4py

    The mpi4py package is used to run domain-decomposed simulations from Python. It is not required for using OpenMOC on a single machine. To install pillow on Ubuntu, issue the following command:

sudo apt-get install python-mpi4py