PLMD::SwitchingFunction Class Reference

Small class to compure switching functions. More...

#include <SwitchingFunction.h>

Public Member Functions
double	calculate (double x, double &df) const
	Compute the switching function. More...
double	calculateSqr (double distance2, double &dfunc) const
	Compute the switching function. More...
std::string	description () const
	Returns a string with a description of the switching function. More...
double	get_d0 () const
	Returns d0. More...
double	get_dmax () const
	Return dmax. More...
double	get_dmax2 () const
	Return dmax squared. More...
double	get_r0 () const
	Returns r0. More...
void	set (int nn, int mm, double r_0, double d_0)
	Set a "rational" switching function. More...
void	set (const std::string &definition, std::string &errormsg)
	Set an arbitrary switching function. More...

Static Public Member Functions
static void	registerKeywords (Keywords &keys)

Private Types
enum	{   rational, exponential, gaussian, smap,   cubic, tanh, cosinus, matheval,   leptontype, nativeq }
	Type of function. More...

Private Member Functions
double	do_rational (double rdist, double &dfunc, int nn, int mm) const
	Low-level tool to compute rational functions. More...

Private Attributes
int	a =0
	Parameters for smap function. More...
int	b =0
double	beta =0.0
double	c =0.0
double	d =0.0
double	d0 =0.0
	Minimum distance (before this, function is one) More...
double	dmax =0.0
	Maximum distance (after this, function is zero) More...
double	dmax_2 =0.0
	Square of dmax, useful in calculateSqr() More...
std::vector< lepton::CompiledExpression >	expression
	Lepton expression. More...
std::vector< lepton::CompiledExpression >	expression_deriv
	Lepton expression for derivative. More...
bool	fastrational =false
	Set to true for fast rational functions (depending on x**2 only) More...
bool	init =false
	This is to check that switching function has been initialized. More...
double	invr0 =0.0
	Inverse of scaling length. More...
double	invr0_2 =0.0
	Square of invr0, useful in calculateSqr() More...
double	lambda =0.0
std::string	lepton_func
	Function for lepton;. More...
std::vector< double * >	lepton_ref
std::vector< double * >	lepton_ref_deriv
bool	leptonx2 =false
	Set to true if lepton only uses x2. More...
int	mm =0
int	nn =6
	Exponents for rational function. More...
double	ref =0.0
double	shift =0.0
double	stretch =1.0
	Parameters for stretching the function to zero at d_max. More...
enum PLMD::SwitchingFunction:: { ... }	type =rational
	Type of function. More...

Detailed Description

Small class to compure switching functions.

Switching functions are created using set() and then can be used with function calculate() or calculateSqr(). Since this is typically computed on a distance vector, the second all (calculateSqr()) allows to skip the calculation of a square root in some case, thus potentially increasing performances.

Member Enumeration Documentation

anonymous enum

anonymous enum

private

Type of function.

Enumerator
rational
exponential
gaussian
smap
cubic
tanh
cosinus
matheval
leptontype
nativeq

Member Function Documentation

calculate()

double PLMD::SwitchingFunction::calculate	(	double	x,
		double &	df
	)		const

Compute the switching function.

Returns s(x). df will be set to the value of the derivative of the switching function _divided_by_x

calculateSqr()

double PLMD::SwitchingFunction::calculateSqr	(	double	distance2,
		double &	dfunc
	)		const

Compute the switching function.

Returns \( s(\sqrt{x})\) . df will be set to the \( \frac{1}{\sqrt{x}}\frac{ds}{d\sqrt{x}}= 2 \frac{ds}{dx}\) (same as calculate()). The advantage is that in some case the expensive square root can be avoided (namely for rational functions, if nn and mm are even and d0 is zero)

description()

std::string PLMD::SwitchingFunction::description

(

)

const

Returns a string with a description of the switching function.

do_rational()

double PLMD::SwitchingFunction::do_rational ( double  rdist, double &  dfunc, int  nn, int  mm  ) const

private

Low-level tool to compute rational functions.

It is separated since it is called both by calculate() and calculateSqr()

get_d0()

double PLMD::SwitchingFunction::get_d0

(

)

const

Returns d0.

get_dmax()

double PLMD::SwitchingFunction::get_dmax

(

)

const

Return dmax.

get_dmax2()

double PLMD::SwitchingFunction::get_dmax2

(

)

const

Return dmax squared.

get_r0()

double PLMD::SwitchingFunction::get_r0

(

)

const

Returns r0.

registerKeywords()

void PLMD::SwitchingFunction::registerKeywords ( Keywords &  keys )

static

set() [1/2]

void PLMD::SwitchingFunction::set	(	int	nn,
		int	mm,
		double	r_0,
		double	d_0
	)

Set a "rational" switching function.

Notice that a d_max is set automatically to a value such that f(d_max)=0.00001.

set() [2/2]

void PLMD::SwitchingFunction::set	(	const std::string &	definition,
		std::string &	errormsg
	)

Set an arbitrary switching function.

Parse the string in definition and possibly returns errors in the errormsg string

Member Data Documentation

a

int PLMD::SwitchingFunction::a =0

private

Parameters for smap function.

b

int PLMD::SwitchingFunction::b =0

private

beta

double PLMD::SwitchingFunction::beta =0.0

private

c

double PLMD::SwitchingFunction::c =0.0

private

d

double PLMD::SwitchingFunction::d =0.0

private

d0

double PLMD::SwitchingFunction::d0 =0.0

private

Minimum distance (before this, function is one)

dmax

double PLMD::SwitchingFunction::dmax =0.0

private

Maximum distance (after this, function is zero)

dmax_2

double PLMD::SwitchingFunction::dmax_2 =0.0

private

Square of dmax, useful in calculateSqr()

expression

std::vector<lepton::CompiledExpression> PLMD::SwitchingFunction::expression

private

Lepton expression.

Warning

Since lepton::CompiledExpression is mutable, a vector is necessary for multithreading!

expression_deriv

std::vector<lepton::CompiledExpression> PLMD::SwitchingFunction::expression_deriv

private

Lepton expression for derivative.

Warning

Since lepton::CompiledExpression is mutable, a vector is necessary for multithreading!

fastrational

bool PLMD::SwitchingFunction::fastrational =false

private

Set to true for fast rational functions (depending on x**2 only)

init

bool PLMD::SwitchingFunction::init =false

private

This is to check that switching function has been initialized.

invr0

double PLMD::SwitchingFunction::invr0 =0.0

private

Inverse of scaling length.

We store the inverse to avoid a division

invr0_2

double PLMD::SwitchingFunction::invr0_2 =0.0

private

Square of invr0, useful in calculateSqr()

lambda

double PLMD::SwitchingFunction::lambda =0.0

private

lepton_func

std::string PLMD::SwitchingFunction::lepton_func

private

Function for lepton;.

lepton_ref

std::vector<double*> PLMD::SwitchingFunction::lepton_ref

private

lepton_ref_deriv

std::vector<double*> PLMD::SwitchingFunction::lepton_ref_deriv

private

leptonx2

bool PLMD::SwitchingFunction::leptonx2 =false

private

Set to true if lepton only uses x2.

mm

int PLMD::SwitchingFunction::mm =0

private

nn

int PLMD::SwitchingFunction::nn =6

private

Exponents for rational function.

ref

double PLMD::SwitchingFunction::ref =0.0

private

shift

double PLMD::SwitchingFunction::shift =0.0

private

stretch

double PLMD::SwitchingFunction::stretch =1.0

private

Parameters for stretching the function to zero at d_max.

type

enum { ... } PLMD::SwitchingFunction::type

Type of function.

---

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

• tools/SwitchingFunction.h
• tools/SwitchingFunction.cpp