Abstract
In this study, we propose a novel algorithm which estimates the optimal division number in the multiple division divide-and-conquer (DCk) for real symmetric tridiagonal eigenproblem. Using the proposed algorithm, we establish an automatically tuned DCk algorithm (ATDCk), in which the quasi-optimal division number used in DCk is automatically determined by a simple but efficient pre-test to the input matrix, prior to the main computation of the eigenvalues and eigenvectors. The efficiency of the ATDCk is confirmed by numerical experiment covering a wide range of test matrices including the physical and statistical models as well as the matrix market. The comparison with the LAPACK routine DSTEVD based on the usual divide-and-conquer with the division number 2 (DC2) is also made.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Cuppen J (1981) A divide and conquer method for the symmetric tridiagonal eigenproblem. Numer Math 36:177–195
Dhillon I (1997) A new O(n 2) algorithm for the symmetric tridiagonal eigenvalue/eigenvector problem. Technical report UCB/CSD-97-971. University of California, Berkeley
Kuwajima Y, Shigehara T (2005) An extension of divide-and-conquer for real symmetric tridiagonal eigenproblem (in Japanese). Trans Jpn Soc Ind Appl Math 15:89–115
Anderson E, et al (1999) LAPACK users’ guide, 2rd edn. Society for Industrial and Applied Mathematics, Philadelphia, PA, USA
Rutter JD (1994) A serial implementation of cuppen’s divide and conquer algorithm for the symmetric eigenvalue problem. Computer Science Division Report, University of California, Berkeley
Arbenz P, Golub GH (1988) On the spectral decomposition of hermitian matrices modified by low rank perturbations with applications. SIAM J Matrix Anal Appl 9:40–58
Kuwajima Y, Shigehara T (2006) An improvement of multiple division divide-and-conquer for real symmetric tridiagonal eigenproblem (in Japanese). Trans Jpn Soc Ind Appl Math 16: 453–480
Bunch JR, Nielsen CP, Sorensen DC (1978) Rank-one modification of the symmetric eigenproblem. Numer Math 31:31–48
Mehta ML (2004) Random matrices, 3rd edn. Pure and applied mathematics (Amsterdam), vol 142. Elsevier/Academic, Amsterdam
Matrix Market, http://math.nist.gov/MatrixMarket
Acknowledgements
This work was partially supported by Grant-in-Aid for Scientific Research (C) No.19560058.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer New York
About this chapter
Cite this chapter
Ishikawa, Y., Tamura, J., Kuwajima, Y., Shigehara, T. (2011). Automatic Tuning of the Division Number in the Multiple Division Divide-and-Conquer for Real Symmetric Eigenproblem. In: Naono, K., Teranishi, K., Cavazos, J., Suda, R. (eds) Software Automatic Tuning. Springer, New York, NY. https://doi.org/10.1007/978-1-4419-6935-4_6
Download citation
DOI: https://doi.org/10.1007/978-1-4419-6935-4_6
Published:
Publisher Name: Springer, New York, NY
Print ISBN: 978-1-4419-6934-7
Online ISBN: 978-1-4419-6935-4
eBook Packages: EngineeringEngineering (R0)