Installation Guide for FEniCSx v0.9.0

Installation Guide for FEniCSx v0.9.0#

This guide provides detailed, copy-paste ready instructions to install FEniCSx v0.9.0 and its dependencies across various operating systems, including dedicated automated workflows for Windows (via WSL 2) and HPC environments (via Spack).

1. Linux Installation (Native or WSL2 Terminal)#

For native Linux systems or when working directly inside a WSL2 distribution’s terminal, you have two primary options: the APT repository for standard users or Spack for advanced/HPC users.

1.1. Local Installation of FEniCSx v0.9.0 (APT Method)#

Note: These instructions are tested for Ubuntu 24.04 LTS. This method installs FEniCSx directly into your system’s packages.

Step 1 — Add the FEniCSx Repository#

Access the package and update your shell environment:

sudo add-apt-repository ppa:fenics-packages/fenics
sudo apt update

Step 2 — Install FEniCSx and Dependencies#

Install FEniCSx at the specific version (0.9.0.1) and all necessary packages for meshing, plotting, and environment management.

# Install FEniCSx v0.9.0
sudo apt install -y 'fenicsx=2:0.9.0.1*'

# Install Python tools, Pandas, Gmsh, and rendering libraries for GUI/plotting
sudo apt install python3-pip python3-venv python3-pandas -y
sudo apt install gmsh xvfb libgl1-mesa-dri libglx-mesa0 libglu1-mesa mesa-utils -y

2. Windows Installation (Automated Batch Scripts)#

For Windows users, using the provided batch scripts is the most reliable way to set up an isolated environment with FEniCSx. Ensure Docker Desktop is installed and running for all Docker methods.

2.2. Option B: WSL2 Local Installer (Native venv in WSL)#

This sets up a dedicated WSL distribution (FEniCSxenv) with a FEniCSx Python Virtual Environment (venv) that auto-activates upon login.

  1. Files: Use the files from the WSL_Local_installer folder (containing install_local_fenicsx.bat, apt-installed-list.txt, etc.).

  2. Installation: Run the installer from Windows (Command Prompt/PowerShell):

    install_local_fenicsx.bat

    You will be prompted to enter the WSL username you created during the initial Ubuntu setup.

  3. Usage: Launch the dedicated environment from Windows:

    wsl -d FEniCSxenv

    The fenicsx-env will be automatically active. Run scripts with:

    python3 filename.py

2.3. Option C: Windows Direct Docker#

This method runs Docker commands directly from Windows, relying on Docker Desktop’s native file mapping.

  1. Prerequisites: Ensure Docker Desktop is running.

  2. Files: Use the files from the Windows_docker_installer folder (containing FEniCSx_docker_0-9-0.bat, etc.).

  3. Image Build: Execute the build script. This will also integrate any custom packages found in the src folder.

    FEniCSx_docker_0-9-0.bat

  4. Usage:

    # Interactive shell
Interactive_FEniCSx_0-9-0.bat

# Non-interactive script run (e.g., Test.py or Main.py)
FEniCSx_AGGREGATE.bat

3. HPC / Advanced Installation (Spack Package Manager)#

Spack is the recommended tool for managing complex dependencies on High-Performance Computing (HPC) clusters, allowing for precise control over compilers, MPI, and linear algebra libraries like PETSc.

Spack Installation Procedure (Detailed)#

This procedure was tested on Red Hat Enterprise Linux 8.10.

Step 1 — Get and Configure Spack#

Download and set up the Spack environment.

# 1. Get Spack v0.23.0
wget [https://github.com/spack/spack/releases/download/v0.23.0/spack-0.23.0.tar.gz](https://github.com/spack/spack/releases/download/v0.23.0/spack-0.23.0.tar.gz)

# 2. Untar the archive
tar -xf spack-0.23.0.tar.gz

# 3. Optional: Disable local config (recommended for multi-user/HPC setups)
export SPACK_DISABLE_LOCAL_CONFIG=true

# 4. Activate Spack
cd spack-0.23.0
source ./share/spack/setup-env.sh

Step 2 — Configure Compilers and Environment#

Add your system compiler to Spack, and optionally install a newer GCC version if needed.

# 1. Add system compiler (must be pre-installed on the OS)
spack compiler find 

# 2. Optional: Install a modern compiler (e.g., GCC 12)
spack install gcc@12
spack compiler find $(spack location -i gcc@12)

Step 3 — Create and Install FEniCSx Environment#

Create a dedicated Spack environment, add the required packages, and initiate the installation.

# 1. Create a Spack environment
spack env create MyFenicsxEnv

# 2. Activate the environment
spack env activate MyFenicsxEnv

# 3. Add Core FEniCSx dependencies and solvers (specify exact versions for reproducibility)
# PETSc config: Disable Trilinos, enable Fortran, Mumps, and SuperLU-Dist
spack add petsc@3.22.1 ~trilinos +fortran +mumps +superlu-dist
spack add python@3.11.9
spack add py-fenics-dolfinx@0.9.0
spack add py-mpi4py@4.0.1
spack add py-numpy@2.1.2
spack add py-matplotlib
spack add py-pandas
spack add py-gmsh

# 4. Install all packages in the environment
spack install

Success! You have installed FEniCSx v0.9.0 and its parallel dependencies.

Step 4 — Usage and Activation#

To use the environment in any new terminal session or submission script:

# 1. Activate Spack environment (add this to your .bashrc or .bash_profile)
source /path/to/spack-0.23.0/share/spack/setup-env.sh

# 2. Activate your custom environment
spack env activate MyFenicsxEnv

# 3. Run a simulation using MPI
mpirun -np 4 python3 my_fenicsx_script.py

HPC Batch Scripts (Slurm): You can reference the file Slurm_spack_fenicsx_0_9_0.sh for an example of how to correctly source and activate the Spack environment within a Slurm job submission script.

Contents