Abstract
We discuss the implementation of the bisection algorithm for the computation of the eigenvalues of symmetric tridiagonal matrices in a context of mixed precision arithmetic. This approach is motivated by the emergence of processors which carry out floating-point operations much faster in single precision than they do in double precision. Perturbation theory results are used to decide when to switch from single to double precision. Numerical examples are presented.
Similar content being viewed by others
References
Anderson, E., et al.: LAPACK Users’ Guide. SIAM, Philadelphia (1999)
Blackford, L., et al.: ScaLAPACK Users’ Guide. SIAM, Philadelphia (1997)
Buttari, A., et al.: Mixed precision iterative refinement techniques for the solution of dense linear systems. Int. J. High Perform. Comput. Appl. 21, 457–466 (2007)
Buttari, A., et al.: Using mixed precision for sparse matrix computations to enhance the performance while achieving 64-bit accuracy. ACM Trans. Math. Softw. TOMS 34(4), 17:1–17:22 (2008)
Demmel, J.W.: The inherent inaccuracy of implicit tridiagonal QR. LAPACK working note 15 (1992)
Demmel, J.W., Li, X.: Faster numerical algorithms via exception handling. IEEE Trans. Comput. 43, 983–992 (1992)
Demmel, J.W., Dhillon, I., Ren, H.: On the correctness of some bisection-like parallel eigenvalue algorithms in floating point arithmetic. Electr. Trans. Numer. Anal. 3, 116–149 (1995)
Demmel, J.W.: Applied Numerical Linear Algebra. SIAM, Philadelphia (1997)
Giraud, l, Haidar, A., Watson, L.: Mixed-precision preconditioners in parallel domain decomposition solvers. Lect. Notes Comput. Sci. Eng 60, 357–364 (2008)
Golub, G., Van Loan, C.: Matrix Computations, 2nd edn. The Johns Hopkins University Press, Baltimore (1989)
Higham, N.: The rise of mixed precision arithmetic. https://nickhigham.wordpress.com/2015/10/20/the-rise-of-mixed-precision-arithmetic/. Accessed Nov 2015
Kurzak, J., Dongarra, J.: Implementation of mixed precision in solving systems of linear equations on the CELL processor. Concurr. Comput. Pract. Exp. 19, 1371–1385 (2007)
Langou, J., et al.: Exploiting the performance of 32 bit floating point arithmetic in obtaining 64 bit accuracy (revisiting iterative refinement for linear systems. In: Proceedings of the ACM/IEEE 2006 Conference. IEEE Computer Society Press (2006)
Parlett, B.N.: The Symmetric Eigenvalue Problem. Prentice-Hall, Englewood Cliffs (1998)
Petschow, M., Quintana-Orti, E.S., Bientinesi, P.: Improved accuracy and parallelism for MRRR-based eigensolvers—a mixed precision approach. SIAM J. Sci. Comput. 36(2), C240–C263 (2014)
Ralha, R.: Perturbation splitting for more accurate eigenvalues. SIAM J. Matrix Anal. Appl. 31, 75–91 (2009)
Volkov, V., Demmel, J.W.: Using GPUs to accelerate the bisection algorithm for finding eigenvalues of symmetric tridiagonal matrices. LAPACK working note 197 (2008)
Wilkinson, J.H.: The Algebraic Eigenvalue Problem. Oxford University Press, London (1965)
Yamazaki, I., Tomov, S., Dongarra, J.: Mixed-precision Choleski QR factorization and its case studies on multicore CPU with multiple GPUs. SIAM J. Sci. Comput. 37(3), C307–C330 (2015)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Ralha, R. Mixed Precision Bisection. Math.Comput.Sci. 12, 173–181 (2018). https://doi.org/10.1007/s11786-018-0336-6
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11786-018-0336-6