Since 2011, I am associate professor at the Universitat Politècnica de Catalunya and adjoint researcher at the International Center for Numerical Methods in Engineering (CIMNE). At CIMNE, I am leading the Large Scale Scientific Computing department, originally supported by the Starting Independent Research Grant COMFUS.
I work on finite element methods, numerical analysis and large scale computing. I am interested in stabilized finite elements and discontinuous Galerkin techniques, mainly focused on fluid problems, e.g., the (Navier-)Stokes equations. I am also interested in the finite element numerical approximation of multiphysics problems and efficient and robust solvers for the resulting linear systems.
In my team, we develop the multiphysics and parallel finite element software FEMPAR, which allows us to simulate different physical problems (fluid mechanics, magnetohydrodynamics, electromagnetics, elasticity, etc.). FEMPAR also includes an extensive numerical linear algebra package. In particular, we have designed an abstract implementation of block (physics-based) preconditioners for multiphysics problems. In order to solve large scale problems, we have developed in FEMPAR a highly scalable implementation of (balancing) domain decomposition techniques. We have proved perfect weak scalability for coercive three-dimensional problems (the Poisson problem and linear elasticity) up to 458,672 cores in JUQUEEN (Germany), the largest supercomputer in Europe. The largest problem solved with FEMPAR up to now involved 30 billion unknowns (November 2014).
Previously, I have worked at the Applied Mathematics Departments at the Politecnico di Milano and Sandia National Labs.