Documentation for module arnoldi_methods

methods for arnoldi iteration

source: arnoldi_methods.F
Loading...

public Subroutines/Functions:

Interface for the initialization of the arnoldi subspace creation currently it can only setup a random vector but can be improved to various types of restarts easily
Interface for the routine calcualting the implicit restarts Currently all based on lapack
here the iterations are performed and the krylov space is constructed
Interface to compute the eigenvalues of a nonsymmetric matrix This is only the serial version
Interface for the initialization of the arnoldi subspace creation for the generalized eigenvalue problem
here the iterations are performed and the krylov space for the generalized eigenvalue probelm is created
in the generalized eigenvalue the matrix depende on the projection therefore the outer loop has to build a new set of matrices for the inner loop

SUBROUTINEarnoldi_init(matrix, vectors, arnoldi_data)

Interface for the initialization of the arnoldi subspace creation currently it can only setup a random vector but can be improved to various types of restarts easily

Arguments:
:: matrix(:) pointer to the matrices as described in main interface
:: vectors work vectors for the matrix vector multiplications
:: arnoldi_data all data concerning the subspace

SUBROUTINEarnoldi_iram(arnoldi_data)

Interface for the routine calcualting the implicit restarts Currently all based on lapack

Arguments:
:: arnoldi_data ...

SUBROUTINEbuild_subspace(matrix, vectors, arnoldi_data)

here the iterations are performed and the krylov space is constructed

Arguments:
:: matrix(:) see above
:: vectors see above
:: arnoldi_data see above

SUBROUTINEcompute_evals(arnoldi_data)

Interface to compute the eigenvalues of a nonsymmetric matrix This is only the serial version

Arguments:
:: arnoldi_data ...

SUBROUTINEgev_arnoldi_init(matrix, matrix_arnoldi, vectors, arnoldi_data)

Interface for the initialization of the arnoldi subspace creation for the generalized eigenvalue problem

Arguments:
:: matrix(:) pointer to the matrices as described in main interface
:: matrix_arnoldi(:) ...
:: vectors work vectors for the matrix vector multiplications
:: arnoldi_data all data concerning the subspace

SUBROUTINEgev_build_subspace(matrix, vectors, arnoldi_data)

here the iterations are performed and the krylov space for the generalized eigenvalue probelm is created

Arguments:
:: matrix(:) see above
:: vectors see above
:: arnoldi_data see above

SUBROUTINEgev_update_data(matrix, matrix_arnoldi, vectors, arnoldi_data)

in the generalized eigenvalue the matrix depende on the projection therefore the outer loop has to build a new set of matrices for the inner loop

Arguments:
:: matrix(:) see above
:: matrix_arnoldi(:) ...
:: vectors ...
:: arnoldi_data see above