AlgorithmParameters

class AlgorithmParameters(density_matrix_method=None, store_grids=None, store_basis_on_grid=None, store_energy_density_matrix=None, scf_restart_step_length=None, use_symmetries=None, reciprocal_space_paw_quantities=None)

Class for representing the parameters for setting solver parameters.

Some parameter defaults are specific for each calculator, see HuckelCalculator, SlaterKosterCalculator, LCAOCalculator, or PlaneWaveCalculator.

Parameters:

  • density_matrix_method (DiagonalizationSolver | FeastSolver | ChebyshevExpansionSolver | PEXSISolver | PlaneWaveResidualMinimizer | GeneralizedDavidsonSolver | PPCGSolver) – The method used for obtaining the density matrix.
    Default: PPCGSolver for PlaneWaveCalculator. DiagonalizationSolver for the other calculators.

  • store_grids (bool) – Enable or disable the storage of grids that can be reused in the SCF process. True will increase the memory usage, but reduce the computation time.
    Default: True

  • store_basis_on_grid (Automatic | True | False) – Control the storage of the basis functions on a real space grid. When enabled, increases the memory usage but reduces the computation time.
    Default: Automatic

  • store_energy_density_matrix (bool | Automatic) – Enable or disable the computation and storage of the energy density matrix while the density matrix is calculated. When enabled, can slow down the SCF loop but speeds up subsequent evaluations of forces and/or stress.
    Default: Disabled unless forces and/or stress will be evaluated immediately after the self-consistent calculation (Automatic).

  • scf_restart_step_length (PhysicalQuantity of type length) – The maximum distance by which an atom can move so that the calculation can be restarted from the previous initial state.
    Default: 0.1 * Ang

  • use_symmetries (Automatic | True | False) – Whether the symmetries of the configuration should be used to reduce the number of k-points. If set to Automatic the symmetries are disabled for a single k-point and enabled for multiple k-points.
    Default: True

  • reciprocal_space_paw_quantities (bool) – If this flag is set to true, the paw quantities will be calculated in reciprocal space, this can only be done when we use periodic boundary conditions.
    Default: True

densityMatrixMethod()
Returns:

The advanced solver.

Return type:

DiagonalizationSolver | FeastSolver | ChebyshevExpansionSolver | PEXSISolver | PlaneWaveResidualMinimizer | GeneralizedDavidsonSolver | PPCGSolver

static enableSymmetries(is_paw, spin_type)

Plane-wave PAW spin-orbit does not yet support the use of symmetries. Enable or disable the option.

WARNING: This emulate the behavior of PlaneWaveCalculator._useSymmetries()

Thus this should change if that change.

Parameters:

  • is_paw (bool) – True, the PlaneWave-PAW is used.

  • spin_type (enum) – The spin type.

Returns:

Truem the use symmetries.

Return type:

bool

static noSCFDensityMatrixMethod()
Returns:

The solver to use for LCAO when no SCF iterations are done.

Return type:

DiagonalizationSolver

scfRestartStepLength()
Returns:

The SCF restart step length of this calculator.

Return type:

PhysicalQuantity of type length

storeBasisOnGrid()
Returns:

True if the storage of the basis functions on a real space grids is enabled

Return type:

bool

storeEnergyDensityMatrix()
Returns:

True if the simultaneous computation and storage of the energy density matrix during the calculation of the density matrix is enabled.

Return type:

bool | Automatic

storeGrids()
Returns:

True if the storage of grids is enabled

Return type:

bool

uniqueString()

Return a unique string representing the state of the object.

useSymmetries()
Returns:

Whether the symmetries of the configuration should be used to reduce the number of k-points.

Return type:

bool

Usage Examples

Setup a calculation that uses the Diagonalization Solver.

density_matrix_method = DiagonalizationSolver()
algorithm_parameters = AlgorithmParameters(density_matrix_method)
calculator = LCAOCalculator(algorithm_parameters=algorithm_parameters)

Setup a calculation that uses the Feast Solver.

density_matrix_method = FeastSolver(lambda_min=-10*Hartree, lambda_max=10*Hartree, number_of_eigenvalues=10)
algorithm_parameters = AlgorithmParameters(density_matrix_method)
calculator = LCAOCalculator(algorithm_parameters=algorithm_parameters)