ESSEX (Equipping Sparse Solvers for Exascale) investigates programming concepts and numerical algorithms for scalable, efficient and robust iterative sparse matrix applications on exascale systems.
The overall aim is to deliver a collection of broadly usable and scalable sparse eigenvalue solvers with high hardware efficiency for the computer architectures to come. Project activities are organized along the traditional software layers of low-level parallel building blocks (kernels), algorithm implementations, and applications. The classic abstraction boundaries separating these layers are broken by strongly integrating objectives: scalability, numerical reliability, fault tolerance, and holistic performance and power engineering.
The departement for High-Performance Computing at the Institute for Software Technology contributes the central integration framework PHIST, including core functionality, linear solvers and the Jacobi-Davidson method. Furthermore, we contribute
to the library of optimized kernels GHOST and the fault
tolerance library CRAFT, developed at Erlangen Regional