DocumentCode :
2633796
Title :
Performance evaluation of a new parallel preconditioner
Author :
Gremban, Keith D. ; Miller, Gary L. ; Zagha, Marco
Author_Institution :
Sch. of Comput. Sci., Carnegie Mellon Univ., Pittsburgh, PA, USA
fYear :
1995
fDate :
25-28 Apr 1995
Firstpage :
65
Lastpage :
69
Abstract :
The linear systems associated with large, sparse, symmetric, positive definite matrices are often solved iteratively using the preconditioned conjugate gradient method. We have developed a new class of preconditioners, support tree preconditioners, that are based on the connectivity of the graphs corresponding to the matrices and are well-structured for parallel implementation. We evaluate the performance of support tree preconditioners by comparing them against two common types of preconditioners: diagonal scaling and incomplete Cholesky. Support tree preconditioners require less overall storage and less work per iteration than incomplete Cholesky preconditioners. In terms of total execution time, support tree preconditioners outperform both diagonal scaling and incomplete Cholesky preconditioners
Keywords :
conjugate gradient methods; parallel processing; software performance evaluation; sparse matrices; trees (mathematics); diagonal scaling preconditioner; graph connectivity; incomplete Cholesky preconditioner; iterative method; linear systems; overall storage; parallel preconditioner; performance evaluation; preconditioned conjugate gradient method; sparse matrices; support tree preconditioners; total execution time; Acceleration; Art; Computational modeling; Computer science; Concurrent computing; Convergence; Gradient methods; Laplace equations; Linear systems; Symmetric matrices;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel Processing Symposium, 1995. Proceedings., 9th International
Conference_Location :
Santa Barbara, CA
Print_ISBN :
0-8186-7074-6
Type :
conf
DOI :
10.1109/IPPS.1995.395915
Filename :
395915
Link To Document :
بازگشت