NTPoly::Matrix_lsr Class Reference

A datatype for storing a CSR matrix (real values). More...

#include <SMatrix.h>

Public Member Functions
	Matrix_lsr (int columns, int rows)
	Matrix_lsr (std::string file_name)
	Matrix_lsr (const NTPoly::TripletList_r &list, int rows, int columns)
	Matrix_lsr (const NTPoly::Matrix_lsr &matB)
int	GetRows () const
	Get the number of rows in a matrix.
int	GetColumns () const
	Get the number of columns in a matrix.
void	ExtractRow (int row_number, Matrix_lsr &row_out) const
void	ExtractColumn (int column_number, Matrix_lsr &column_out) const
void	Scale (double constant)
void	Increment (const NTPoly::Matrix_lsr &matB, double alpha, double threshold)
double	Dot (const NTPoly::Matrix_lsr &matB) const
void	PairwiseMultiply (const NTPoly::Matrix_lsr &matA, const NTPoly::Matrix_lsr &matB)
void	Gemm (const NTPoly::Matrix_lsr &matA, const NTPoly::Matrix_lsr &matB, bool isATransposed, bool isBTransposed, double alpha, double beta, double threshold, NTPoly::MatrixMemoryPool_r &memory_pool)
void	Transpose (const NTPoly::Matrix_lsr &matA)
	Transpose a sparse matrix.
void	Print () const
	Print the sparse matrix to the console.
void	WriteToMatrixMarket (std::string file_name) const
void	MatrixToTripletList (NTPoly::TripletList_r &triplet_list) const
	~Matrix_lsr ()
	Standard destructor.

Detailed Description

A datatype for storing a CSR matrix (real values).

Constructor & Destructor Documentation

Matrix_lsr() [1/4]

NTPoly::Matrix_lsr::Matrix_lsr	(	int	columns,
		int	rows
	)

Basic constructor.

Parameters

columns	number of columns for the matrix.
rows	number of rows for the matrix.

Matrix_lsr() [2/4]

NTPoly::Matrix_lsr::Matrix_lsr

(

std::string

file_name

)

Construct from a matrix market file.

Parameters

file_name

matrix market file name.

Matrix_lsr() [3/4]

NTPoly::Matrix_lsr::Matrix_lsr	(	const NTPoly::TripletList_r &	list,
		int	rows,
		int	columns
	)

Construct from a triplet list.

Parameters

list	a list of triplet values to set in the matrix.
columns	number of columns for the matrix.
rows	number of rows for the matrix.

Matrix_lsr() [4/4]

NTPoly::Matrix_lsr::Matrix_lsr

(

const NTPoly::Matrix_lsr &

matB

)

Copy constructor.

Parameters

matB	the matrix to copy from.

Member Function Documentation

Dot()

double NTPoly::Matrix_lsr::Dot

(

const NTPoly::Matrix_lsr &

matB

)

const

Matrix dot product.

Parameters

matB	matrix to dot with.

Returns

the dot product of this and matB.

ExtractColumn()

void NTPoly::Matrix_lsr::ExtractColumn	(	int	column_number,
		Matrix_lsr &	column_out
	)		const

Extract a column from the matrix into the compressed vector representation

Parameters

column_number	the column to extract
column_out	the matrix representing that column

ExtractRow()

void NTPoly::Matrix_lsr::ExtractRow	(	int	row_number,
		Matrix_lsr &	row_out
	)		const

Extract a row from the matrix into the compressed vector representation.

Parameters

row_number	the row to extract
row_out	the matrix representing that row

Gemm()

void NTPoly::Matrix_lsr::Gemm	(	const NTPoly::Matrix_lsr &	matA,
		const NTPoly::Matrix_lsr &	matB,
		bool	isATransposed,
		bool	isBTransposed,
		double	alpha,
		double	beta,
		double	threshold,
		NTPoly::MatrixMemoryPool_r &	memory_pool
	)

This := alpha*matA*op( matB ) + beta*this

Parameters

matA
matB
isATransposed	true if A is already transposed.
isBTransposed	true if B is already transposed.
alpha	scaling value.
beta	scaling value.
threshold	for flushing small values.
memory_pool	a memory pool to use for storing intermediates.

Increment()

void NTPoly::Matrix_lsr::Increment	(	const NTPoly::Matrix_lsr &	matB,
		double	alpha,
		double	threshold
	)

This = alpha*MatrixB + This(AXPY).

Parameters

matB	matrix to add.
alpha	scale for the matrix.
threshold	for flushing small values.

MatrixToTripletList()

void NTPoly::Matrix_lsr::MatrixToTripletList

(

NTPoly::TripletList_r &

triplet_list

)

const

Compute a triplet list from the entries in a matrix.

Parameters

triplet_list

output.

PairwiseMultiply()

void NTPoly::Matrix_lsr::PairwiseMultiply	(	const NTPoly::Matrix_lsr &	matA,
		const NTPoly::Matrix_lsr &	matB
	)

Pairwise multiply two sparse matrices.

Parameters

matA
matB

Scale()

void NTPoly::Matrix_lsr::Scale

(

double

constant

)

Scale the matrix by a constant.

Parameters

constant

value to scale by.

WriteToMatrixMarket()

void NTPoly::Matrix_lsr::WriteToMatrixMarket

(

std::string

file_name

)

const

Write the sparse matrix to file.

Parameters

file_name

file to print to.

---

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

• /Users/wddawson/Documents/NTPoly/NTPoly-Max/Source/CPlusPlus/SMatrix.h
• /Users/wddawson/Documents/NTPoly/NTPoly-Max/Source/CPlusPlus/SMatrix.cc