powerit Namespace Reference

Namespace for power iterations. More...

Classes
struct	Options
	Options for compute(). More...
struct	Result
	Result of compute(). More...

Functions
template<typename Data_ , class Engine_ >
void	fill_starting_vector (size_t order, Data_ *vector, Engine_ &engine)
template<class Multiply_ , typename Data_ >
Result< Data_ >	compute_core (size_t order, Multiply_ multiply, Data_ *vector, const Options &opt)
template<typename Task_ , class Run_ >
void	parallelize (int num_workers, Task_ num_tasks, Run_ run_task_range)
template<typename Data_ , class Engine_ >
Result< Data_ >	compute (size_t order, const Data_ *matrix, bool row_major, Data_ *vector, Engine_ &engine, const Options &opt)
template<typename Data_ >
Result< Data_ >	compute (size_t order, const Data_ *matrix, bool row_major, Data_ *vector, const Options &opt)

Detailed Description

Namespace for power iterations.

Function Documentation

compute() [1/2]

template<typename Data_ >

Result< Data_ > powerit::compute	(	size_t	order,
		const Data_ *	matrix,
		bool	row_major,
		Data_ *	vector,
		const Options &	opt
	)

Perform power iterations on an array containing a diagonizable matrix. This overload assumes that a random starting vector has already been generated.

Template Parameters

Data_

Floating-point type for the data.

Parameters

	order	Order of the square matrix.
[in]	matrix	Pointer to an array containing an order-by-order diagonalizable matrix.
	row_major	Whether matrix is row-major.
[in,out]	vector	Pointer to an array of length order. On input, this should contain a random starting vector. On output, this contains the estimate of the first eigenvector.
	opt	Further options.

Returns

Result containing the first eigenvalue and other diagnostics.

compute() [2/2]

template<typename Data_ , class Engine_ >

Result< Data_ > powerit::compute	(	size_t	order,
		const Data_ *	matrix,
		bool	row_major,
		Data_ *	vector,
		Engine_ &	engine,
		const Options &	opt
	)

Perform power iterations on a diagonizable matrix to find the first eigenvalue/vector. This overload generates a starting vector from an existing (P)RNG.

Template Parameters

Data_	Floating-point type for the data.
Engine_	Any C++11-compliant random number generator class.

Parameters

	order	Order of the square matrix.
[in]	matrix	Pointer to an array containing an order-by-order diagonalizable matrix.
	row_major	Whether matrix is row-major.
[out]	vector	Pointer to an array of length order. On output, this contains the estimate of the first eigenvector.
	engine	Instance of the random number generator.
	opt	Further options.

Returns

Result containing the first eigenvalue and other diagnostics.

compute_core()

template<class Multiply_ , typename Data_ >

Result< Data_ > powerit::compute_core	(	size_t	order,
		Multiply_	multiply,
		Data_ *	vector,
		const Options &	opt
	)

Perform power iterations on some diagonizable matrix to find the first eigenvalue/vector. This overload generates a starting vector from an existing (P)RNG.

Template Parameters

Multiply_	Function that performs a matrix dot product. This is used to abstract away the representation of the actual matrix.
Data_	Floating-point type for the data.
Engine_	Any C++11-compliant random number generator class.

Parameters

	order	Order of the square matrix.
	multiply	Function to perform a matrix dot product with the working eigenvector. This should accept buffer, a std::vector<Data_>& of length order; and vector, a const Data_* as described below. It should fill buffer with the product of the matrix and vector. The return value is ignored.
[out]	vector	Pointer to an array of length order. On output, this contains the estimate of the first eigenvector.
	opt	Further options.

Returns

Result containing the first eigenvalue and other diagnostics.

fill_starting_vector()

template<typename Data_ , class Engine_ >

void powerit::fill_starting_vector	(	size_t	order,
		Data_ *	vector,
		Engine_ &	engine
	)

Template Parameters

Data_	Floating-point type for the data.
Engine_	Any C++11-compliant random number generator class.

Parameters

	order	Length of the array in vector.
[out]	vector	Pointer to an array of length order. On output, this will be filled with draws from a standard normal distribution.
	engine	Instance of the random number generator.

parallelize()

template<typename Task_ , class Run_ >

void powerit::parallelize	(	int	num_workers,
		Task_	num_tasks,
		Run_	run_task_range
	)

Template Parameters

Task_	Integer type for the number of tasks.
Run_	Function to execute a range of tasks.

Parameters

num_workers	Number of workers.
num_tasks	Number of tasks.
run_task_range	Function to iterate over a range of tasks within a worker.

By default, this is an alias to subpar::parallelize_range(). However, if the POWERIT_CUSTOM_PARALLEL function-like macro is defined, it is called instead. Any user-defined macro should accept the same arguments as subpar::parallelize_range().