Parallel implementation of a multiblock method with approximate subdomain solution

From MaRDI portal





Investigation of the parallelized iterative solution of linear systems resulting from the discretization of a Poisson type partial differential equation. Domain decomposition is used, resulting in a block diagonal system where the main diagonal blocks are solved in an inner iteration as preconditioner for the outer GCR iteration. Interesting theoretical (model) and practical investigations on a workstation cluster and on a Cray T3E are made with different inner solvers: RILUD is the best one.











This page was built for publication: Parallel implementation of a multiblock method with approximate subdomain solution

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1294610)