Recommendation 0. Read the Installation Guide

OpenMC has an excellent installation guide that covers the dependencies. Several practices, discussed below, may make the whole process easier, cleaner, and more self-contained.

Recommendation 1. Set up a Conda Environment

When building OpenMC from source, using a conda environment to install dependencies allows different versions of various dependencies to co-exist on the same machine. It also eases installation of OpenMC’s Python API. To create an environment with the name openmc-env, run:

conda create -n openmc-env

Recommendation 2. Install Dependencies

Install all the dependencies listed in the installation guide using conda in your new environment. For example, to install the Python dependencies, run:

conda install numpy scipy pandas pytest matplotlib lxml
uncertainties

This should install the required compilers and HDF5 libraries needed for OpenMC.

[!NOTE] The recommendations for installing h5py in development mode are in the installation guide. If you chose to do this, you should set HDF5_DIR=$CONDA_PREFIX (for building h5py) and HDF5_ROOT=$CONDA_PREFIX(for building OpenMC).