The School aims at sharing models and numerical methods used by the state-of-the-art CFD research and industrial codes running on High Performance Computing (HPC) clusters, relying on the expertise of internationally recognized scholars and HPC specialists.
The School is based on the four consecutive successful editions of the virtual winter school, with an increasing interest and number of students every year.
The lectures will present challenges involved with solving nonlinear PDEs in problems related to fluid dynamics, using state-of-the-art algorithms. Efficient implementation on massively parallel clusters will be especially targeted.
The lectures will provide a step-by-step didactic introduction to the numerical methods, each lecture being followed by tutorial, hands-on sessions dedicated to efficient parallel implementation.
Introductory and advanced notions about parallelization strategies (MPI, OpenMPI), use of Graphic Accelerators, as well as plug-in numerical libraries, will also be provided.