molpro::linalg Namespace Reference

Namespaces
namespace	array
namespace	iterativesolver
namespace	itsolv
	4-parameter interpolation of a 1-dimensional function given two points for which function values and first derivatives are known.

Classes
struct	is_complex
struct	is_complex< std::complex< T > >
struct	real_type
	The real type underlying a (possibly complex) scalar type. More...
struct	real_type< std::complex< T > >

Typedefs
template<typename T >
using	real_type_t = typename real_type< T >::type
	The real type underlying T, i.e. T itself for a real type and U for std::complex<U>.

Functions
const std::shared_ptr< const molpro::Options >	options ()
	Get the Options object associated with iterative-solver.
void	set_options (const molpro::Options &options)
	Set the options for iterative-solver.
template<typename T >
T	conjugate (const T &x)
	Complex conjugate, staying within the scalar type; the identity for a real type.
template<typename T >
real_type_t< T >	real_part (const T &x)
	The real part of a scalar; the value itself for a real type.
template<typename T >
real_type_t< T >	imaginary_part (const T &x)
	The imaginary part of a scalar; zero for a real type.
template<typename value_type >
real_type_t< value_type >	precision_scaled (double tolerance_for_double)
	Rescale a tolerance that was calibrated for IEEE double precision to the working precision.

Typedef Documentation

real_type_t

template<typename T >

using molpro::linalg::real_type_t = typedef typename real_type<T>::type

The real type underlying T, i.e. T itself for a real type and U for std::complex<U>.

Quantities that are real by construction – eigenvalues of a hermitian matrix, singular values, norms and the thresholds they are compared against – are expressed in this type rather than in the scalar type of the problem.

Function Documentation

conjugate()

template<typename T >

T molpro::linalg::conjugate

(

const T &

x

)

Complex conjugate, staying within the scalar type; the identity for a real type.

imaginary_part()

template<typename T >

real_type_t< T > molpro::linalg::imaginary_part

(

const T &

x

)

The imaginary part of a scalar; zero for a real type.

options()

const std::shared_ptr< const molpro::Options > molpro::linalg::options

(

)

Get the Options object associated with iterative-solver.

Returns

The current option set. If none has been assigned, in the Molpro context the "ITERATIVE-SOLVER" option set is used, otherwise empty.

precision_scaled()

template<typename value_type >

real_type_t< value_type > molpro::linalg::precision_scaled

(

double

tolerance_for_double

)

Rescale a tolerance that was calibrated for IEEE double precision to the working precision.

A tolerance of the form "this quantity vanishes to within rounding error" is a fixed multiple of the machine epsilon of the arithmetic used. The constants in this library were chosen for double precision; this function preserves their safety margin, measured in machine epsilons, at any other precision, and returns the constant unchanged when the working precision is double.

real_part()

template<typename T >

real_type_t< T > molpro::linalg::real_part

(

const T &

x

)

The real part of a scalar; the value itself for a real type.

set_options()

void molpro::linalg::set_options

(

const molpro::Options &

options

)

Set the options for iterative-solver.

Parameters

options