Advertisement

Numerische Mathematik

, Volume 58, Issue 1, pp 779–805 | Cite as

A parallel algorithm for the eigenvalues and eigenvectors of a general complex matrix

  • Gautam M. Shroff
Article

Summary

A new parallel Jacobi-like algorithm is developed for computing the eigenvalues of a general complex matrix. Most parallel methods for this problem typically display only linear convergence, Sequential ‘norm-reducing’ algorithms also exist and they display quadratic convergence in most cases. The new algorithm is a parallel form of the ‘norm-reducing’ algorithm due to Eberlein. It is proven that the asymptotic convergence rate of this algorithm is quadratic. Numerical experiments are presented which demonstrate the quadratic convergence of the algorithm and certain situations where the convergence is slow are also identified. The algorithm promises to be very competitive on a variety of parallel architectures. In particular, the algorithm can be implemented usingn2/4 processors, takingO(n log2n) time for random matrices.

Subject classification

AMS (MOS) CR O65 
This is a preview of subscription content, log in to check access.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Bischof, C.: Computing the singular value decomposition on a distributed system of vector processors. Technical report 87 869, Department of Computer Science, Cornell University 1987Google Scholar
  2. 2.
    Brent, R.P., Luk, F.T.: The solution of singular-value and symmetric eigenvalue problems on multiprocessor arrays. SIAM J. Sci. Stat. Comput.6, 69–84 (1985)Google Scholar
  3. 3.
    Causey, R.L.: Computing eigenvalues of non hermitian matrices by methods of Jacobi type. J. SIAM6, 172–181 (1958)Google Scholar
  4. 4.
    Eberlein, P.J.: A Jacobi method for the automatic computation of eigenvalues and eigenvectors of an arbitrary matrix. J. SIAM10, 74–88 (1962)Google Scholar
  5. 5.
    Eberlein, P.J.: On the Schur decomposition of a matrix for parallel computation. IEEE Trans. Comput.36, 167–174 (1987)Google Scholar
  6. 6.
    Eberlein, P.J., Moothroyd, J.: Solution to the eigenproblem by a normreducing Jacobi-type method. Numer. Math.4, 24–40 (1968)Google Scholar
  7. 7.
    Fan, K., Hoffman, A.J.: Lower bounds for the rank and location of eigenvalues of a matrix. National Bureau of Standards Applied Math. Series39, 117–130 (1954)Google Scholar
  8. 8.
    Forsythe, G.E., Henrici, P.: The cyclic Jacobi method for computing the principal values of a complex matrix. Trans. Am. Math. Soc.94, 1–23 (1960)Google Scholar
  9. 9.
    Goldstine, H.H., Horwitz, L.P.: A procedure for the diagonalization of normal matrices. J ACM6, 176–195 (1959)Google Scholar
  10. 10.
    Goldstine, H.H., Murray, F.J., Neumann, J. von: The Jacobi method for real symmetric matrices. J ACM6, 59–96 (1959)Google Scholar
  11. 11.
    Golub, G., Van Loan, C.: Matrix computations. Johns Hopkins University Press (1983)Google Scholar
  12. 12.
    Hansen, E.R.: On Jacobi methods and block Jacobi methods for computing matrix eigenvalues. PhD thesis, Stanford University 1960Google Scholar
  13. 13.
    Jacobi, C.G.J.: Über ein Verfahren die in Theorie der Säkularstörungen vorkommenden Gleichungen numerisch aufzulösen. J. Reine Angew. Math.30, 51–95 (1846)Google Scholar
  14. 14.
    Luk, F.T., Park, H.T.: On parallel Jacobi orderings. Technical report EE-CEG-86-5, School of Electrical Engineering, Cornell University 1986Google Scholar
  15. 15.
    Luk, F.T., Park, H.T.: A proof of convergence for two parallel Jacobi SVD algorithms. Technical Report EE-CEG-86-12, School of Electrical Engineering, Cornell University 1986Google Scholar
  16. 16.
    Osborne, E.E.: On pre-conditioning of matrices. J. ACM7, 338–345 (1960)Google Scholar
  17. 17.
    Pardekooper, M.H.C.: An eigenvalue algorithm based on norm-reducing transformations. PhD thesis, Technische Universiteit Eindhoven 1969Google Scholar
  18. 18.
    Pardekooper, M.H.C.: A quadratically convergent parallel Jacobi process for diagonally dominant matrices with distinct eigenvalues. J. Comput. Appl. Math.27, 3–16 (1989)Google Scholar
  19. 19.
    Ruhe, A.: On the quadratic convergence of a generalization of the Jacobi method to arbitrary matrices. BIT8, 210–231 (1968)Google Scholar
  20. 20.
    Rutishauser, H.: Une methode pour le calcul des values propres des matrices nonsymetriques. Comptes Rendus259, 2758 (1964)Google Scholar
  21. 21.
    Sameh, A.H.: On Jacobi and Jacobi-like algorithms for a parallel computer. Math. Comput.25, 579–590 (1971)Google Scholar
  22. 22.
    Schreiber, R.: Solving eigenvalue and singular value problems on an undersized systolic array. SIAM J. Sci. Stat. Comput.7, 441–451 (1986)Google Scholar
  23. 23.
    Shroff, G.: Parallel Jacobi algorithms for the algebraic eigenvalue problem. PhD thesis, Rensselaer Polytechnic Institute 1990Google Scholar
  24. 24.
    Shroff, G., Schreiber, R.: On the convergence of the cyclic Jacobi method for parallel block orderings. SIAM J. Matrix Anal. Appl.10 (1989)Google Scholar
  25. 25.
    Stewart, G.W.: A Jacobi-like algorithm for computing the schur decomposition of a nonhermitian matrix. SIAM J. Sci. Stat. Comput.6, 853–864 (1985)Google Scholar
  26. 26.
    Veselic, K.: On a class of Jacobi-like procedures for diagonalizing arbitrary real matrices. Numer. Math.33, 157–172 (1979)Google Scholar
  27. 27.
    Veselic, K.: A quadratically convergence Jacobi-like method for real matrices with complex conjugate eigenvalues. Numer. Math.33, 425–435 (1979)Google Scholar
  28. 28.
    Wilkinson, J.H.: Almost diagonal matrices with multiple or close eigenvalues. Linear Algebra Appl.1, 1–12 (1968)Google Scholar

Copyright information

© Springer-Verlag 1991

Authors and Affiliations

  • Gautam M. Shroff
    • 1
  1. 1.Computer Science DepartmentRensselaer Polytechnic InstituteTroyUSA

Personalised recommendations

Cite article

Buy options