Abstract
The need for block formulations of numerical algorithms in order to achieve high performance rates on pipelined vector supercomputers with hierarchical memory systems is discussed. Methods for analysing the performance of standard kernel linear algebra subroutines, such as the BLAS (levels 1, 2 and 3), are reviewed. We extend the L 2 block Gram-Schmidt algorithm of Jalby and Philippe to orthogonalization in elliptic norms induced by the inner product u TM u, where M = B T B is a positive definite matrix. A performance analysis of this block algorithm is presented along with experimental results obtained on the Cray-2 supercomputer.
Preview
Unable to display preview. Download preview PDF.
References
K. GALLIVAN, W. JALBY, and U. MEIER, “The use of BLAS3 in linear algebra on a parallel processor with a hierarchical memory”, SIAM J. Sci. Stat. Comput., 8 (1987), pp. 1079–84.
K. GALLIVAN, R. J. PLEMMONS, and A. SAMEH, “Parallel algorithms for dense linear algebra computations”, SIAM Review, 32 (1990), pp. 54–135.
K. GALLIVAN, W. JALBY, U. MEIER, and A. SAMEH, “The impact of hierarchical memory systems on linear algebra algorithm design”, Intl. J. Supercomputer Appl., 2 (1987), pp. 1079–1084.
W. JALBY and U. MEIER, “Optimizing matrix operations on a parallel multiprocessor with a hierarchical memory system”, ICPP Proceedings, pp. 429–32, 1986.
W. JALBY and B. PHILIPPE, “Stability analysis and improvement of the block Gram-Schmidt algorithm”, SIAM J. Sci. Stat. Comput., 6 (1991), pp. 26–50.
S. J. THOMAS and R. V. M. ZAHAR, “Efficient Orthogonalization in the M-norm”, Congressus Numerantium, vol. 80, 1990.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1992 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Thomas, S.J. (1992). A block algorithm for orthogonalization in elliptic norms. In: Bougé, L., Cosnard, M., Robert, Y., Trystram, D. (eds) Parallel Processing: CONPAR 92—VAPP V. VAPP CONPAR 1992 1992. Lecture Notes in Computer Science, vol 634. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-55895-0_434
Download citation
DOI: https://doi.org/10.1007/3-540-55895-0_434
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-55895-8
Online ISBN: 978-3-540-47306-0
eBook Packages: Springer Book Archive