Hierarchical auto-correlation function algorithm. More...

#include <AutoCorrStage.h>

Inheritance diagram for Util::AutoCorrStage< Data, Product >:

Public Member Functions
	AutoCorrStage ()
	Constructor. More...

virtual	~AutoCorrStage ()
	Destructor. More...

void	setParam (int bufferCapacity=64, int maxStageId=0, int blockFactor=2)
	Set all parameters and allocate to initialize state. More...

virtual void	sample (Data value)
	Sample a value. More...

void	clear ()
	Clear accumulators and destroy descendants. More...

template<class Archive >
void	serialize (Archive &ar, const unsigned int version)
	Serialize to/from an archive. More...

Accessors
int	bufferCapacity_
	Physical capacity (# of elements) of buffer, corr, and nCorr. More...

int	maxStageId_
	Maximum allowed stage index (controls maximum degree of blocking). More...

int	blockFactor_
	Number of values per block (ratio of intervals for successive stages). More...

void	output (std::ostream &out)
	Output the autocorrelation function, assuming zero mean. More...

void	output (std::ostream &out, Product aveSq)
	Output the autocorrelation function. More...

int	bufferCapacity () const
	Return capacity of history buffer. More...

int	bufferSize () const
	Return current size of history buffer. More...

long	nSample () const
	Return the number of sampled values. More...

long	stageInterval () const
	Return the number of primary values per block at this stage. More...

Product	autoCorrelation (int t) const
	Return autocorrelation at a given time, assuming zero average. More...

Product	autoCorrelation (int t, Product aveSq) const
	Return autocorrelation at a given lag time. More...

double	corrTime () const
	Estimate of autocorrelation time, in samples. More...

double	corrTime (Product aveSq) const
	Numerical integration of autocorrelation function. More...

void	allocate ()
	Allocate memory and initialize to empty state. More...

bool	hasChild () const
	Does this have a child AutoCorrStage? More...

AutoCorrStage &	child ()
	Return the child AutoCorrStage by reference. More...

virtual void	registerDescendant (AutoCorrStage< Data, Product > *ptr)
	Register the creation of a descendant stage. More...

template<class Archive >
void	serializePrivate (Archive &ar, const unsigned int version)
	Serialize private data members, and descendants. More...

Detailed Description

template<typename Data, typename Product>
class Util::AutoCorrStage< Data, Product >

Hierarchical auto-correlation function algorithm.

This class calculates an autocorrelation function for a sequence x(i) of values of a variable or object of type Data. The resulting autocorrelation function is and array of values of type Product, where C(j) = <x(i-j), x(i)>. Here <A,B> denotes an inner product of type Product for objects A and B of type Data.

The meaning of the inner product is defined for various data types b the overloaded function Product product(Data, Data) that is defined for double, complex and Vector data in the product.h file.

The zero value for variables of type Data is returned by the overloaded function void setToZero(Data) method defined in the setToData.h file.

This class implements a hierarchical algorithm to calculate C(j). The algorithm is implemented by a linked list of AutoCorrStage objects. Each object in this list is assigned an integer chainId.
The "primary" AutoCorrStage object in this list, with chainId=0, calculates the autocorrelation for a primary sequence of primary Data values that are passed to the sample method of this object. For each n > 0, the object with chainId = n calculates the autocorrelation function for a sequence of values in which each value is an average of a block of blockFactor**n consecutive values of the primary sequence or, equivalently, an average of blockFactor consecutive values of the sequence maintained by the parent object with chainId = n-1. Additional stages are added to this list dynamically as needed.

Definition at line 53 of file AutoCorrStage.h.

Constructor & Destructor Documentation

◆ AutoCorrStage()

template<typename Data , typename Product >

Util::AutoCorrStage< Data, Product >::AutoCorrStage

Constructor.

This constructor creates a primary AutoCorrStage object with stageId = 0 and stageInterval = 1. A private constructor is used to recursively create descendant stages as needed.

Definition at line 30 of file AutoCorrStage.tpp.

References Util::setToZero().

◆ ~AutoCorrStage()

template<typename Data , typename Product >

Util::AutoCorrStage< Data, Product >::~AutoCorrStage

virtual

Destructor.

Recursively destroy all descendant stages.

Definition at line 79 of file AutoCorrStage.tpp.

Member Function Documentation

◆ setParam()

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::setParam	(	int	bufferCapacity = `64`,
		int	maxStageId = `0`,
		int	blockFactor = `2`
	)

Set all parameters and allocate to initialize state.

Parameters

bufferCapacity	max. number of values stored in buffer
maxStageId	maximum stage index (0=primary)
blockFactor	ratio of block sizes of subsequent stages

Definition at line 90 of file AutoCorrStage.tpp.

◆ sample()

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::sample ( Data value )

virtual

Sample a value.

Parameters

value current Data value

Definition at line 121 of file AutoCorrStage.tpp.

References Util::product(), and Util::setToZero().

◆ clear()

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::clear

Clear accumulators and destroy descendants.

Definition at line 103 of file AutoCorrStage.tpp.

References Util::setToZero().

Referenced by Util::AutoCorrStage< Data, Product >::allocate().

◆ serialize()

template<typename Data , typename Product >

template<class Archive >

void Util::AutoCorrStage< Data, Product >::serialize	(	Archive &	ar,
		const unsigned int	version
	)

Serialize to/from an archive.

Parameters

ar	archive
version	archive version id

Definition at line 168 of file AutoCorrStage.tpp.

◆ output() [1/2]

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::output ( std::ostream & out )

Output the autocorrelation function, assuming zero mean.

This calls output(std::ostream out, Product aveSq) with a zero value for aveSq.

Parameters

out	output stream.

Definition at line 259 of file AutoCorrStage.tpp.

References Util::setToZero().

◆ output() [2/2]

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::output	(	std::ostream &	out,
		Product	aveSq
	)

Output the autocorrelation function.

The parameter avSq = ave(x)^2 is subtracted from the correlation function ave(x(t)x(0)).

Parameters

out	output stream
aveSq	square of ave(x)

Definition at line 270 of file AutoCorrStage.tpp.

◆ bufferCapacity()

template<typename Data , typename Product >

int Util::AutoCorrStage< Data, Product >::bufferCapacity

Return capacity of history buffer.

Definition at line 231 of file AutoCorrStage.tpp.

◆ bufferSize()

template<typename Data , typename Product >

int Util::AutoCorrStage< Data, Product >::bufferSize

Return current size of history buffer.

Definition at line 238 of file AutoCorrStage.tpp.

Referenced by Util::AutoCorrelation< Data, Product >::maxDelay().

◆ nSample()

template<typename Data , typename Product >

long Util::AutoCorrStage< Data, Product >::nSample

Return the number of sampled values.

Definition at line 245 of file AutoCorrStage.tpp.

◆ stageInterval()

template<typename Data , typename Product >

long Util::AutoCorrStage< Data, Product >::stageInterval

Return the number of primary values per block at this stage.

Definition at line 252 of file AutoCorrStage.tpp.

Referenced by Util::AutoCorrelation< Data, Product >::maxDelay().

◆ autoCorrelation() [1/2]

template<typename Data , typename Product >

Product Util::AutoCorrStage< Data, Product >::autoCorrelation ( int t ) const

Return autocorrelation at a given time, assuming zero average.

This calls autoCorrelations(t, aveSq) with a zero value for for aveSq.

Parameters

t	the lag time, in Data samples

Definition at line 296 of file AutoCorrStage.tpp.

References Util::setToZero().

◆ autoCorrelation() [2/2]

template<typename Data , typename Product >

Product Util::AutoCorrStage< Data, Product >::autoCorrelation	(	int	t,
		Product	aveSq
	)		const

Return autocorrelation at a given lag time.

The parameter aveSq is subtracted from ave(x(t)x(0)) in output.

Parameters

t	the lag time, in Data samples
aveSq	square ave(x(t))

Definition at line 307 of file AutoCorrStage.tpp.

◆ corrTime() [1/2]

template<typename Data , typename Product >

double Util::AutoCorrStage< Data, Product >::corrTime

Estimate of autocorrelation time, in samples.

This variant assumes a zero average.

Definition at line 319 of file AutoCorrStage.tpp.

References Util::setToZero().

◆ corrTime() [2/2]

template<typename Data , typename Product >

double Util::AutoCorrStage< Data, Product >::corrTime ( Product aveSq ) const

Numerical integration of autocorrelation function.

This function returns the time integral of the autocorrelation function. The parameter aveSq is subtracted from ave(x(t)x(0)) in the integrand.

Parameters

aveSq square ave(x(t))

Definition at line 330 of file AutoCorrStage.tpp.

References Util::setToZero().

◆ allocate()

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::allocate

protected

Allocate memory and initialize to empty state.

Definition at line 357 of file AutoCorrStage.tpp.

References Util::AutoCorrStage< Data, Product >::clear().

◆ hasChild()

template<typename Data , typename Product >

bool Util::AutoCorrStage< Data, Product >::hasChild

inlineprotected

Does this have a child AutoCorrStage?

Definition at line 305 of file AutoCorrStage.h.

◆ child()

template<typename Data , typename Product >

AutoCorrStage< Data, Product > & Util::AutoCorrStage< Data, Product >::child

inlineprotected

Return the child AutoCorrStage by reference.

Definition at line 312 of file AutoCorrStage.h.

◆ registerDescendant()

template<typename Data , typename Product >

void Util::AutoCorrStage< Data, Product >::registerDescendant ( AutoCorrStage< Data, Product > * ptr )

protectedvirtual

Register the creation of a descendant stage.

This should be called only by a root stage.

Parameters

ptr	pointer to a descendant AutoCorrStage.

Reimplemented in Util::AutoCorrelation< Data, Product >.

Definition at line 385 of file AutoCorrStage.tpp.

◆ serializePrivate()

template<typename Data , typename Product >

template<class Archive >

void Util::AutoCorrStage< Data, Product >::serializePrivate	(	Archive &	ar,
		const unsigned int	version
	)

protected

Serialize private data members, and descendants.

Parameters

ar	archive
version	archive version id

Definition at line 186 of file AutoCorrStage.tpp.

Member Data Documentation

◆ bufferCapacity_

template<typename Data , typename Product >

int Util::AutoCorrStage< Data, Product >::bufferCapacity_

protected

Physical capacity (# of elements) of buffer, corr, and nCorr.

Definition at line 193 of file AutoCorrStage.h.

◆ maxStageId_

template<typename Data , typename Product >

int Util::AutoCorrStage< Data, Product >::maxStageId_

protected

Maximum allowed stage index (controls maximum degree of blocking).

Definition at line 196 of file AutoCorrStage.h.

◆ blockFactor_

template<typename Data , typename Product >

int Util::AutoCorrStage< Data, Product >::blockFactor_

protected

Number of values per block (ratio of intervals for successive stages).

Definition at line 199 of file AutoCorrStage.h.

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

Public Member Functions

Accessors

Detailed Description

Constructor & Destructor Documentation

◆ AutoCorrStage()

◆ ~AutoCorrStage()

Member Function Documentation

◆ setParam()

◆ sample()

◆ clear()

◆ serialize()

◆ output() [1/2]

◆ output() [2/2]

◆ bufferCapacity()

◆ bufferSize()

◆ nSample()

◆ stageInterval()

◆ autoCorrelation() [1/2]

◆ autoCorrelation() [2/2]

◆ corrTime() [1/2]

◆ corrTime() [2/2]

◆ allocate()

◆ hasChild()

◆ child()

◆ registerDescendant()

◆ serializePrivate()

Member Data Documentation

◆ bufferCapacity_

◆ maxStageId_

◆ blockFactor_