Template Class InfNorm

Defined in File InfNorm.hpp

Nested Relationships

Nested Types

Struct InfNorm::Parameters

Inheritance Relationships

Base Type

public AnalysisManager::NativeDynamicSolver::ErrorNorm< ScalarT, IdxT > (Template Class ErrorNorm)

Class Documentation

template<class ScalarT, typename IdxT> class InfNorm : public AnalysisManager::NativeDynamicSolver::ErrorNorm<ScalarT, IdxT>

The infinity error norm, which requires the error in every component in the system to meet tolerance.

Public Functions

virtual RealT errorNorm(State &err, State &y, State &yprev, GridKit::LinearAlgebra::VectorHandler<ScalarT, IdxT> &handler, GridKit::memory::MemorySpace memspace) const final

Calculate an error to be used by a step controller. Typically, an error > 1 indicates an error which does not meet tolerances, while an error < 1 indicates an error which meets tolerances. For that reason, tolerances should be included in the calculation of the error.

Todo:: Allow this method to fail, since it will likely involve linear algebra calls.

Parameters:

err – The state error residual being measured.
y – The state that the error was calculated from. Can be used for proper relative error normalization.
yprev – The state from the previous step. Can be used for proper relative error normalization.
handler – A vector handler which can be used to facilitate vector operations.
memspace – The memory space which vector operations should be performed in/.

Returns:

The error.

Public Members

std::unique_ptr<State> out_: The final vector which will have its norm taken.

std::unique_ptr<State> scale_: The vector which will be used to scale the error in accordance with the tolerances.

std::unique_ptr<State> yprev_abs_: The absolute value of yprev. Used to calculate scale_.

struct Parameters

The configurable parameters of the error norm.

Public Members

std::unique_ptr<State> abs_tol_: A vector of absolute tolerances for each component. The norm will attempt to reject errors in each component above this tolerance.

RealT rel_tol_: The relative tolerance. The norm will attempt to reject any error larger in percentage of the solution’s maximum element than this.