Equipping Sparse Solvers for Exascale - A Survey of the DFG Project ESSEX

Abstract

The ESSEX project investigates computational issues arising at exascale for large-scale sparse eigenvalue problems and develops programming concepts and numerical methods for their solution. The project pursues a coherent co-design of all software layers where a holistic performance engineering process guides code development across the classic boundaries of application, numerical method and basic kernel library. Within ESSEX the numerical methods cover both widely applicable solvers such as classic Krylov, Jacobi-Davidson or recent FEAST methods as well as domain specific iterative schemes relevant for the ESSEX quantum physics application. This presentation introduces the project structure and presents selected results which demonstrate the potential impact of ESSEX for efficient sparse solvers on highly scalable heterogeneous supercomputers.

In the second project phase from 2016 to 2018, the ESSEX consortium will include partners from the Universities of Tokyo and of Tsukuba. Extensions of existing work will regard numerically reliable computing methods, scalability improvements by leveraging functional parallelism in asynchronous preconditioners, hiding and reducing communication cost, improving load balancing by advanced partitioning schemes, as well as the treatment of non-Hermitian matrix problems.