An efficient and accurate parallel algorithm for the singular value problem of bidiagonal matrices (Q1347046)

From MaRDI portal
scientific article
Language Label Description Also known as
English
An efficient and accurate parallel algorithm for the singular value problem of bidiagonal matrices
scientific article

    Statements

    An efficient and accurate parallel algorithm for the singular value problem of bidiagonal matrices (English)
    0 references
    0 references
    0 references
    0 references
    2 April 1995
    0 references
    A divide and conquer type algorithm for computing the eigenvalues of a symmetric tridiagonal matrix is described. It uses the eigenvalues of the two diagonal block submatrices as starting values in a Laguerre iteration, which can be performed in parallel. This so called split merge algorithm is applied to bidiagonal matrices to compute the singular value decomposition. Several numerical comparisons to standard algorithms are presented.
    0 references
    parallel computation
    0 references
    divide and conquer type algorithm
    0 references
    eigenvalues
    0 references
    symmetric tridiagonal matrix
    0 references
    Laguerre iteration
    0 references
    split merge algorithm
    0 references
    bidiagonal matrices
    0 references
    singular value decomposition
    0 references
    numerical comparisons
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references