Abstract
For the solutions of linear systems of equations with unsymmetric coefficient matrices, we has proposed an improved version of the quasi-minimal residual (IQMR) method by using the Lanczos process as a major component combining elements of numerical stability and parallel algorithm design. For Lanczos process, stability is obtained by a couple two-term procedure that generates Lanczos vectors scaled to unit length. The algorithm is derived such that all inner products and matrix-vector multiplications of a single iteration step are independent and communication time required for inner product can be overlapped efficiently with computation time. Therefore, the cost of global communication on parallel distributed memory computers can be significantly reduced. In this paper, we describe an efficient implementation of this method which is particularly well suited to problems with irregular sparsity pattern. The corresponding communication cost is independent of the sparsity pattern with several performance improvement techniques such as overlapping computation and communication, balancing the computational load. The performance is demonstrated by numerical experimental results carried out on massively parallel distributed memory computer Parsytec GC/Power Plus.
Preview
Unable to display preview. Download preview PDF.
References
H. M. Bucker. Isoefficiciency analysis of parallel QMR-like iterative methods and its implications on parallel algorithm design. Technical Report KFA-ZAM-IB-9604, Central Institute for Applied Mathematics, Research Centre Julich, Germany, January 1996.
H. M. Bucker and M. Sauren. A parallel version of the quasi-minimal residual method based on coupled two-term recurrences. In Proceedings of Workshop on Applied Parallel Computing in Industrial Problems and Optimization (Para96). Technical University of Denmark, Lyngby, Denmark, Springer-Verlag, August 1996.
H. M. Bucker and M. Sauren. A parallel version of the unsymmetric Lanczos algorithm and its application to QMR. Technical Report KFA-ZAM-IB-9606, Central Institute for Applied Mathematics, Research Centre Julich, Germany, March 1996.
E. de Sturler. A parallel variant of the GMRES(m). In Proceedings of the 13th IMACS World Congress on Computational and Applied Mathematics. IMACS, Criterion Press, 1991.
E. de Sturler and H. A. van der Vorst. Reducing the effect of the global communication in GMRES(m) and CG on parallel distributed memory computers. Technical Report 832, Mathematical Institute, University of Utrecht, Utrecht, The Netherland, 1994.
J. J. Dongarra, I. S. Duff, D. C. Sorensen, and H. A. van der Vorst. Solving Linear Systems on Vector and Shared Memory Computers. SIAM, Philadelphia, PA, 1991.
R. W. Freund, M. H. Gutknecht, and N. M. Nachtigal. An implementation of the look-ahead Lanczos algorithm for non-Hermitian matrices. SIAM Journal on Scientific and Statistical Computing, 14:137–158, 1993.
R. W. Freund and N. M. Nachtigal. QMR: a quasi-minimal residual method for non-Hermitian linear systems. Numerische Mathematik, 60:315–339, 1991.
R. W. Freund and N. M. Nachtigal. An implementation of the QMR method based on coupled two-term recurrences. SIAM Journal on Scientific and Statistical Computing, 15(2):313–337, 1994.
B. Hendrickson, R. Leland, and S. Plimpton. An efficient parallel algorithm for matrix-vector multiplication. International Journal of High Speed Computing, 7(1):73–88, 1995.
V. Kumar, A. Grama, A. Gupta, and G. Karypis. Introduction to Parallel Computing: Design and Analysis of Algorithms. Benjamin/Cummings, Redwood City, 1994.
A. T. Ogielski and W. Aiello. Sparse matrix computations on parallel processor arrays. SIAM Journal on Scientific and Statistical Computing, 14:519–530, 1993.
T. Yang and H. X. Lin. The improved quasi-minimal residual method on massively distributed memory computers. In Proceedings of The International Conference on High Performance Computing and Networking (HPCN-97), April 1997.
T. Yang and H. X. Lin. Isoefficiency analysis of the improved quasi-minimal residual method on massively distributed memory computers. Submitted to The 2rd International Conference on Parallel Processing and Applied Mathmetics (PPAM-97), September 1997.
T. Yang and H. X. Lin. Performance evaluation of the improved quasi-minimal residual method on massively distributed memory computers. Submitted to 5th International Conference on Applications of High-Performance Computers in Engineering (AHPCE-97), July 1997.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Yang, T., Lin, HX. (1997). Efficient implementation of the improved quasi-minimal residual method on massively distributed memory computers. In: Bilardi, G., Ferreira, A., Lüling, R., Rolim, J. (eds) Solving Irregularly Structured Problems in Parallel. IRREGULAR 1997. Lecture Notes in Computer Science, vol 1253. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-63138-0_8
Download citation
DOI: https://doi.org/10.1007/3-540-63138-0_8
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63138-5
Online ISBN: 978-3-540-69157-0
eBook Packages: Springer Book Archive