Simulators for Exascale Computations
This transversal work package aims at introducing innovative computational algorithms well suited to the numerical simulation of complex phenomena on exascale architectures.
The ultimate goal is to set up a computational platform on top of new numerical and computational techniques driven by the challenge of handling complex problems of billions degrees of freedoms in a minute of time. In order to achieve such an outstanding performance, cutting-edge researches contribute to improve each one of the four main aspects in computational simulators, namely, (1) numerical schemes for PDEs (Partial Differential Equations), (2) sparse linear solvers, (3) adaptivity, and (4) data management. Particularly the first topic is concerned with scalable implementations of high order schemes for wave propagation models. The second topic will develop and demonstrate the benefits of generic (i.e. algebraic) parallel solvers for large sparse linear systems of equations. The third topic addresses mesh and (local) time-step adaptive algorithms in order to optimize the use of computational resources. The fourth topic focuses on leveraging techniques to support simulation data management as required by the pre- and pos-processing steps involved in highly complex simulations.
EU coordinator: INRIA. Brazil coordinator: LNCC