Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Computational Science, Engineering & Technology Series
DEVELOPMENTS IN PARALLEL, DISTRIBUTED, GRID AND CLOUD COMPUTING FOR ENGINEERING
Edited by: B.H.V. Topping and P. Iványi
Asynchronous Schwarz Methods for Peta and Exascale Computing
Laboratoire de Mathématiques Appliquées aux Systèmes, Ecole Centrale Paris, Châtenay-Malabry, France
F. Magoulès, "Asynchronous Schwarz Methods for Peta and Exascale Computing", in B.H.V. Topping and P. Iványi, (Editor), "Developments in Parallel, Distributed, Grid and Cloud Computing for Engineering", Saxe-Coburg Publications, Stirlingshire, UK, Chapter 10, pp 229-247, 2013. doi:10.4203/csets.31.10
Keywords: Schwarz algorithm, domain decomposition, interface condition, parallel computing, asynchronous iteration.
Optimized Schwarz methods are very efficient iterative algorithms for the parallel solution of computational mechanics problems. These methods are similar to the classical Schwarz methods, but they use absorbing interface conditions between the subdomains. These interface conditions are then optimized for efficiency and lead to a fast and robust convergence behavior of the iterative algorithm. A wide range of optimization techniques have been proposed and developed in the literature. In this chapter, optimized Schwarz methods are extended to an asynchronous context. The word asynchronous here is related to the iterations of the mathematical algorithm. Numerical experiments illustrate the robustness and efficiency of this original approach when solving the Helmholtz equation on parallel computers.
purchase the full-text of this chapter (price £20)