sfepy.solvers.eigen module

class sfepy.solvers.eigen.LOBPCGEigenvalueSolver(conf, **kwargs)

SciPy-based LOBPCG solver for sparse symmetric problems.

name = 'eig.scipy_lobpcg'

static process_conf(conf, kwargs)

Missing items are set to default values.

Example configuration, all items:

solver_2 = {
    'name' : 'lobpcg',
    'kind' : 'eig.scipy_lobpcg',

    'i_max' : 20,
    'n_eigs' : 5,
    'eps_a' : None,
    'largest' : True,
    'precond' : None,
    'verbosity' : 0,
}

class sfepy.solvers.eigen.PysparseEigenvalueSolver(conf, **kwargs)

Pysparse-based eigenvalue solver for sparse symmetric problems.

name = 'eig.pysparse'

static process_conf(conf, kwargs)

Missing items are set to default values.

Example configuration, all items:

solver_2 = {
    'name' : 'eigen1',
    'kind' : 'eig.pysparse',

    'i_max' : 150,
    'eps_a' : 1e-5,
    'tau' : -10.0,
    'method' : 'qmrs',
    'verbosity' : 0,
    'strategy' : 1,
}

class sfepy.solvers.eigen.ScipyEigenvalueSolver(conf, **kwargs)

SciPy-based solver for both dense and sparse problems (if n_eigs is given).

name = 'eig.scipy'

class sfepy.solvers.eigen.ScipySGEigenvalueSolver(conf, **kwargs)

SciPy-based solver for dense symmetric problems.

name = 'eig.sgscipy'

static process_conf(conf, kwargs)

Missing items are set to default values.

Example configuration, all items:

solver_20 = {
    'name' : 'eigen2',
    'kind' : 'eig.sgscipy',

    'force_n_eigs' : True,
}

sfepy.solvers.eigen.eig(mtx_a, mtx_b=None, n_eigs=None, eigenvectors=True, return_time=None, method='eig.scipy', **ckwargs)

Utility function that constructs an eigenvalue solver given by method, calls it and returns solution.

sfepy.solvers.eigen.standard_call(call)

Decorator handling argument preparation and timing for eigensolvers.