NESO-Tokamak

A NESO-based solver for modelling MAST-U.

Building

The easiest way to build the app is to use the Spack package manager. The build process has been tested with Spack v0.19.0; later versions may also work, but aren't recommended.

1. Install Spack v0.19.0, via the official instructions. On Debian-based systems (e.g. Ubuntu) the following should work:

# Ensure all prerequisites are installed
apt update
apt install -y build-essential ca-certificates coreutils curl environment-modules gfortran git gpg lsb-release python3 python3-distutils python3-venv unzip zip

git clone -c feature.manyFiles=true -b v0.19.0 https://github.com/spack/spack.git $HOME/.spack

# Modify .bashrc such that spack is always initialised in new shells
echo 'export SPACK_ROOT=$HOME/.spack' >> $HOME/.bashrc
echo 'source $SPACK_ROOT/share/spack/setup-env.sh' >> $HOME/.bashrc
export SPACK_ROOT=$HOME/.spack
source $SPACK_ROOT/share/spack/setup-env.sh

2. Next, having git-cloned this repository, change to the root directory (containing this readme) and obtain the required versions of NESO and its dependencies with:

git submodule update --init --recursive

3. The app can then be built with:

spack env activate ./spack -p
spack install -j [np]

where [np] is the number of parallel processes to use in the build. Note that some of the dependencies (particularly nektar++) can take some time to install.

The solver executable will be generated in spack-build-<hash>, where <hash>`` is an 8 character hash generated by Spack. To rebuild the app after modifying the source code, repeat the spack install with the spack environment active.

Running examples

Configuration files for 2D and 3D examples (with and without impurity tracer particles) can be found in the ./examples directory. An easy way to run the examples is (from the top level of the repository):

# Load the mpich spack module
spack load mpich
# Set up and run an example via the helper script
./scripts/run_eg.sh [example_name]

• This will copy ./examples/[example_name] to ./runs/[example_name] and run the solver in that directory with 4 MPI processes by default.
• To run with a different number of MPI processes, use <-n num_MPI>
• The solver executable is assumed to be in the most recently modified spack-build* directory. To choose a different build location, use <-b build_dir_path>.

To change the number of openMP threads used by each process, use

# Run an example with OMP and MPI
OMP_NUM_THREADS=[nthreads] ./scripts/run_eg.sh [example_name]

Postprocessing

• Particle output is generated in an hdf5 file. See particle_velhist.py for an example of reading and postprocessing (requires these packages).
• Fluid output is generated as nektar++ checkpoint files. One way to visualise is them is to convert to .vtu using Nektar's FieldConvert tool and then use Paraview:

# Convert to vtu using FieldConvert (requires xml files as args)
cd runs/[example_name]
spack load nektar
FieldConvert-rg [config_xml] [mesh_xml] [chk_name] [vtu_name]
# e.g. FieldConvert-rg hw.xml square.xml hw_100.chk hw_100.vtu

Known issues / annoyances

• Particle hdf5 output isn't finalised until the solver finishes, so interrupted runs will generate an unreadable file.
• If all fluid boundary conditions are periodic, the nektar++ routine used to solve for the potential arbitrarily sets one of the degrees of freedom to zero in order to avoid a singular solution. This leads to an artifact in the potential (which eventually creates similar artifacts in the vorticity and density fields, see below).

The nektar++ developers are currently working on a fix for this.

---