ePubs

The open archive for STFC research publications

Full Record Details

Persistent URL http://purl.org/net/epubs/work/33115
Record Status Checked
Record Id 33115
Title Parallel algorithms for general sparse systems
Contributors
Abstract We discuss several algorithms for the solution of sparse linear systems on parallel architectures. We will be concerned primarily with shared memory machines and will consider both direct and iterative methods of solution. The direct methods are based on Gaussian elimination and exploit parallelism coming from both the sparsity and from a full matrix kernel. The iterative method is a block projection method using a direct solver within the blocks. We then discuss the development of two codes, one for the solution of general sparse equations, the other for sparse symmetric indefinite systems. Some results on vector multiprocessor supercomputers are presented to illustrate the behaviour of our approaches.
Organisation CSE , CSE-NAG , SERC
Keywords
Funding Information
Related Research Object(s):
Licence Information:
Language English (EN)
Type Details URI(s) Local file(s) Year
Book Chapter or Section In Computer Algorithms for Solving Linear Algebraic Systems: The State of the Art. NATO ASI Series Vol.77 edited by Spedicato E, 277-297. Springer, 1991. https://doi.org/1…978-3-642-76717-3_13 1991