Accession Number : ADP002616

Title :   State of the Art in Eigenvalue Computations,

Corporate Author : CALIFORNIA UNIV BERKELEY

Personal Author(s) : Parlett,B. N.

Report Date : 1983

Pagination or Media Count : 4

Abstract : The subject must be divided up according to two obvious but unappreciated criteria. Matrices are either large or small and the eigenvalues are either real or complex. Moreover, the property that produces real eigenvalues in the overwhelming number of cases is the symmetry of a real matrix. Definition: An n x n matrix is small (in a given computer system) if it and its matrix of eigenvectors can be held as conventional square arrays in the fast (random access) store or memory. Otherwise it is large.

Descriptors :   *Eigenvalues, *Computations, *State of the Art, Eigenvectors, Matrices(Mathematics), Parallel processing

Distribution Statement : APPROVED FOR PUBLIC RELEASE