THE PROJECT

FEMPAR is a scientific software for the simulation of problems governed by partial differential equations (PDEs). It provides a set of state-of-the-art numerical discretizations, including finite element methods, discontinuous Galerkin methods, XFEM, and spline-based functional spaces. The library was originally designed to efficiently exploit distributed-memory supercomputers and easily handle multiphysics problems. It also provides a set of highly scalable numerical linear algebra solvers based on multilevel domain decomposition for the systems of equations that arise from PDE discretizations. Some applications of FEMPAR include the simulation of metal additive manufacturing processes, superconductor devices, breeding blankets in fusion reactors, and nuclear waste repositories.

The FEMPAR project was launched in 2011 as an in-house code by Santiago Badia, Alberto F. Martín, and Javier Principe, at CIMNE and UPC, in the frame of the Starting Grant of Santiago Badia. Six years later, we decided to distribute the code to external users and potential contributors as open source software.

FEMPAR has attained perfect weak scalability for up to 458,672 cores in JUQUEEN (Germany), the largest supercomputer in Europe, solving up to 60 billion unknowns (November 2014). It motivated the inclusion of FEMPAR in the High-Q club of the most scalable European codes.


Institutions

       


Funding

FEMPAR has been funded, directly or indirectly, through different projects and fellowships from the following institutions: