818 Highly Scalable Solvers for Computational PDEs

Hiroshi Kawai, Toyo University
Santiago Badia, Universitat Politècnica de Catalunya
Paul Lin, Sandia National Laboratory
Masao Ogino, Nagoya University
Shioya Ryuuji, Toyo University
Efficient computational solution of high fidelity large-scale problems in computational science and engineering is still a major challenge. Complex applications include difficulties such as transient problems with widely varying time and spatial scales, strongly coupled multiphysics, heterogenous media, nonlinearlities, etc. The development of efficient linear system solvers for these classes of problems has many challenges, especially for high fidelity large-scale simulations, which makes proper preconditioning critical. This minisymposium will focus on highly scalable preconditioners, for example multigrid or domain decomposition approaches, multiphysics solvers (e.g. based on block factorization techniques), nonlinear preconditioning, multiscale solvers for heterogenous problems or space-time solvers. Contributions discussing algorithms that can exploit many-core processors and accelerators are also welcomed.