Demeter¹ is a modern and open source deterministic neutron transport code for reactor simulations. The code is designed to perform 3D fixe source and eigenvalue calculations by solving the multi-group neutron transport equation.

The physics is coded in C++, with Python as user interface for easy and fast usage.

_{1. DEterMinistic nEutron Transport solvER}

• Setup Cmake.
• Add and test Eigen.
• Setup directory structure.
• Code and test Material class.
• Collect source and compile.
• Setup nanobind and write Python bindings for Material.
• Write a python test for Material class.
• Code and test Geometry class.
  • Cell
  • Lattice.
• Code and test angular quadrature
  • Gauss-Legendre
• Code and test Solver module.
  • Compute and update fission source.
  • Compute and update keff.
  • Compute and update scattering source.
  • Base solver.
  • Mesh.
  • Boundary conditions.
• Test Geometry and solver in python.
  • Geometry bindings.
  • Solver bindings.
  • Test
• Write full benchmark
  • In C++.
  • In Python.
• Mesh refiner, rings and sectors for cell
• Enable users to test the lib with codespace and notebooks.
• GPU offload.

• Add AddressSanitizer, ThreadSanitizer, MemorySanitizer.
• Setup Github Action.
• Improve Github Action, by fixing OpenMP not found and with recent version of compiler and Eigen.
• Setup clang-format with Google style.
• Setup clang-tidy.
• Fix how to launch clang-tidy locally and on action.
• Move bindings elsewhere
• Doc style (eg doxygen) and the way to generate it automatically.
• Choose how to handle asserts: thow user errors at runtime and assert internal error on debug build
• Add a Logger (spdlog, or nanolog or fmtlog) and std::format.
• Add a timer.
• Use doctest for unit tests in c++ and python.
• Maybe switch to include/demeter, src, src/binding, test, extra directory structure
• Add -Werror flag
• Setup gcc analizer.
• Generate .clagn-tidy with --warnings-as-errors=
• Peform cland-tidy in github action.

• A C++ compiler compatible with at least version 17 of the standard, eg:
  • GCC 8 and later apt install build-essential,
  • Clang 8 and later apt install clang libomp-dev clang-tidy clang-format.
• CMake 3.4 and newer apt install cmake.
• Eigen 3.3 and newer apt install libeigen3-dev.
• Python 3.8 and newer apt install libpython3-dev
• Doxygen 1.12 and newer apt install doxygen

cd <project-directory>
python -m venv pyenv
source pyenv/bin/activate
pip install numpy matplotlib
pip install .

Following this, you should be able to access the extension

>>> from demeter import *
>>> help(Material)

git clone --recursive https://github.com/kennethassogba/Demeter
mkdir build && cd build
cmake ..
make -j

Then try

./main

cmake -S . -B build/ -D CMAKE_BUILD_TYPE=Debug
cmake --build build/ # build the binaries

cmake -S . -B build/ -D CMAKE_CXX_COMPILER=clang++-18

cmake -S . -B build/ -D CMAKE_CXX_FLAGS=-fsanitize=memory

The documentation is automatically generated using Doxygen. To install it, refer to: https://www.doxygen.nl/manual/install.html

doxygen Doxyfile

You can modify the configuration file Doxyfile to generate the documentation according to your preferences.