The OpenSn Discrete-Ordinates Code

OpenSn is an open-source, massively parallel, radiation transport code designed to solve the discrete-ordinates Boltzmann transport equation for neutral particles in a deterministic manner. OpenSn tackles a variety of problems, including steady-state and dynamical source-driven scenarios, as well as eigenvalue (criticality) problems.

Written in modern C++, OpenSn features a Python interface, making it versatile and accessible. From a simple laptop to a supercomputer, OpenSn can be compiled and executed across a wide range of computing platforms.

OpenSn aggregates decades of research and development in numerical methods and parallel algorithms applied to radiation transport. The most prominent features of OpenSn include:

  • A robust spatial finite-element method that is fully compatible with arbitrary polyhedral cells and adaptively refined grids.

  • Standard angular quadratures, as well as locally refined finite-element angular quadrature rules for the angular variable. Note that uncollided-flux treatment techniques to mitigate ray effects typically found in angular collocation methods will become available at a later date.

  • The traditional multigroup approximation applied to the energy variable. It is worth noting that multi-particle (e.g., neutron+gamma) problems can be seamlessly handled.

  • Advanced iterative solution algorithms are employed to efficiently solve the resulting system of equations, including transport-sweep preconditioning with various synthetic accelerators.

  • Execution of massively parallel simulations on arbitrarily partitioned computational domains, utilizing efficient data aggregation and workflows to maximize performance. The code scales efficiently on a large number of MPI ranks, supporting up to 100,000+ processes.

Recommended publication for citing

Auhtors, “OpenSn: title,” journal, volume, page–page (year).

Contents