Krylov implementation optimised for tensor network applications where calculating \( x^T \cdot A \cdot y \) does not require a matrix-vector application. More...

#include <tn_krylov.h>

Inheritance diagram for syten::Krylov::TensorNetworks< Matrix, Vector, OrthoPolicy, HEffPolicy, SolverPolicy, ApplicationPolicy >:

Collaboration diagram for syten::Krylov::TensorNetworks< Matrix, Vector, OrthoPolicy, HEffPolicy, SolverPolicy, ApplicationPolicy >:

Public Types
using	Matrix_type = Matrix
	type of the supplied matrix More...

using	Vector_type = Vector
	type of the supplied vector More...

Public Member Functions
template<typename... Args>
IterativeExit	solve (IterativeConfig const &conf, Vector &result, Args...args)
	Solves the problem once the input vector is known. More...

template<typename... Args>
auto	solve (Vector &v, Args...args)
	Solves the problem using only the currently existing Krylov space. More...

template<typename... Args>
IterativeExit	solve (Vector v, IterativeConfig const &conf, Vector &result, Args...args)
	Initialises the starting vector and solves the problem. More...

template<typename... Args>
std::tuple< IterativeExit, std::vector< Vector >, DenseTensor< 2, SDef > >	solve_kss (Vector v, IterativeConfig const &conf, Vector &result, Args...args)

	TensorNetworks ()=default
	Default ctor. More...

	TensorNetworks (Matrix m)
	Standard ctor: Pass in the matrix to be solved. More...

template<typename... Params>
void	wrapup (Params &&... params)
	Hook to evaluate data at completion of the current Krylov space calculation. More...

Private Member Functions
bool	solveAtThisStep (IterativeConfig const &conf)
	Returns true if one should attempt a solution at this step, depending on the configuration. More...

Private Attributes
Matrix	H
	The input matrix to be solved. More...

SRDef	initialNorm {1.}
	Norm of the input vector stored here. More...

std::vector< Vector >	krylov
	Set of Krylov vectors constituting the Krylov subspace. More...

Detailed Description

template<typename Matrix, typename Vector, typename OrthoPolicy, typename HEffPolicy, typename SolverPolicy, typename ApplicationPolicy = TensorNetworksImpl::ApplyThenOrtho>
class syten::Krylov::TensorNetworks< Matrix, Vector, OrthoPolicy, HEffPolicy, SolverPolicy, ApplicationPolicy >

Krylov implementation optimised for tensor network applications where calculating \( x^T \cdot A \cdot y \) does not require a matrix-vector application.

Template Parameters

Matrix	type of the input matrix
Vector	type of the input vector
OrthoPolicy	policy class used to orthogonalise the Krylov space
HEffPolicy	policy class which returns the projected effective matrix
SolverPolicy	policy class which tests for convergence and returns the solution once the projected effective matrix is known.

The documentation for this class was generated from the following file:

inc/alg/tn_krylov.h

Public Types

Public Member Functions

Private Member Functions

Private Attributes

Detailed Description