The purpose of this class is to study HPC architectures and some state-of-the-art parallel programming environments, while focusing on performance. The main objective is to decrypt the complexity of nowadays supercomputer architectures and how to exploit them at their full potential, studying practical examples of parallel programs featuring up-to-date parallel programming environments like OpenMP 4.0. This class also gives insight on research projets in the field of parallel programming.Contact François BROQUEDIS
Somes courses on parallel architectures and the way we program them + presentation of state-of-the-art research projects on parallel programming focusing on performance. Students have to present a scientific publication of their choosing to the rest of the class.