Topic: A numerical solver as a business example in HPC
Speaker: Horst Schwichtenberg & Fabian Nick (SCAI)
HPC center: Fraunhofer SCAI
Category of Best Practice: Technological
The HPC Center at Fraunhofer SCAI in cooperation with the other departments of the institute offers a broad range of services (e.g. consulting, training, code optimizing & software development). Moreover, the institute offers licensed HPC ready products. One of the products is the industry-proven and serviced software called SAMG (Algebraic Multigrid Methods for Systems).
In many applications of numerical simulation, for example in fluid flow and structural mechanics, structures and geometries are discretised by means of complex grids. Grids with a finer resolution lead to more accurate simulations. On the other hand, this also increases the size of the system of discretized equations which needs to be solved numerically. Due to the accuracy of simulation results required nowadays, the amount of time needed to solve these systems of equations has become a critical factor. Classical numerical solvers typically do not have the capability to solve such large systems of equations in an acceptable amount of time.
SAMG is a library of subroutines for the highly efficient solution of large linear systems of equations with sparse matrices.
Compared to classical methods (e.g., the ILU-preconditioned conjugate gradient method), SAMG has the advantage of being almost unconditionally numerically scalable. This means that the computational cost using SAMG depends only linearly on the number of unknowns. Depending on the application and problem size, the computational cost can be reduced by one to two orders of magnitude. SAMG can be incorporated into an existing software package as easily as any classical method.
Fact Sheet: A numerical solver as a business example in HPC