molpro::linalg::itsolv::subspace Namespace Reference

Namespaces
namespace	qspace
namespace	util
namespace	xspace

Classes
struct	Dimensions
class	DSpace
	D space stores complement of P+Q+R subspace that is necessary to reproduce the previous solutions. More...
struct	ISubspaceSolver
	Manages solution of the subspace problem and storage of those solutions. More...
class	IXSpace
	Full subspace. More...
class	Matrix
	Matrix container that allows simple data access, slicing, copying and resizing without loosing data. More...
class	PSpace
struct	QSpace
	Container storing the Q space parameters. More...
class	SubspaceSolverDIIS
	Solves non-linear equations using DIIS. More...
class	SubspaceSolverLinEig
	Solves subspace problem for linear eigenvalues and system of linear equations. More...
class	SubspaceSolverOptBFGS
	Solves subspace problem for minimisation using the Steepest Descent algorithm. More...
class	SubspaceSolverOptSD
	Solves subspace problem for minimisation using the Steepest Descent algorithm. More...
class	SubspaceSolverRSPT
class	XSpace

Typedefs
template<typename T = double>
using	SubspaceData = std::map< EqnData, Matrix< T > >
	Equation data blocks of the subspace, in the scalar type of the problem.

Enumerations
enum class	EqnData {   H ,   S ,   rhs ,   value ,   M }
	Kinds of equation data in the subspace. More...

Functions
template<class ML , class MR >
void	transpose_copy (ML &&ml, const MR &mr)
template<class ML , class MR >
void	conjugate_transpose_copy (ML &&ml, const MR &mr)
	Copies the conjugate transpose of mr into ml.
template<class Mat >
std::string	as_string (const Mat &m, int precision=6)
template<typename T >
std::vector< EqnData >	square_blocks (const SubspaceData< T > &data)
	The square matrices, H, S and, if present, M, among the equation data.
template<typename T = double, EqnData... DataTypes>
auto	null_data ()

Typedef Documentation

SubspaceData

template<typename T = double>

using molpro::linalg::itsolv::subspace::SubspaceData = typedef std::map<EqnData, Matrix<T> >

Equation data blocks of the subspace, in the scalar type of the problem.

Enumeration Type Documentation

EqnData

enum class molpro::linalg::itsolv::subspace::EqnData

strong

Kinds of equation data in the subspace.

H is the kernel matrix and S the overlap of the subspace vectors. For a generalised eigenproblem, M is the metric matrix, which then replaces S in the subspace eigenproblem; S remains the overlap used for managing the subspace.

Enumerator
H
S
rhs
value
M

Function Documentation

as_string()

template<class Mat >

std::string molpro::linalg::itsolv::subspace::as_string	(	const Mat &	m,
		int	precision = 6
	)

conjugate_transpose_copy()

template<class ML , class MR >

void molpro::linalg::itsolv::subspace::conjugate_transpose_copy	(	ML &&	ml,
		const MR &	mr
	)

Copies the conjugate transpose of mr into ml.

This is what relates the two off-diagonal blocks of a hermitian matrix; for a real element type it is the same as transpose_copy().

null_data()

template<typename T = double, EqnData... DataTypes>

auto molpro::linalg::itsolv::subspace::null_data

(

)

square_blocks()

template<typename T >

std::vector< EqnData > molpro::linalg::itsolv::subspace::square_blocks

(

const SubspaceData< T > &

data

)

The square matrices, H, S and, if present, M, among the equation data.

transpose_copy()

template<class ML , class MR >

void molpro::linalg::itsolv::subspace::transpose_copy	(	ML &&	ml,
		const MR &	mr
	)