Sparse Matrix Constructors

Thse constructors build an HODLR matrix from a preallocated sparse matrix. More...

Functions
	HODLR_Matrix::HODLR_Matrix (Eigen::SparseMatrix< double > &inputMatrix, int inputSizeThreshold=30, std::string LR_Method="PS_Sparse")
	This constructor initializes the class with a sparse matrix and optional leaf size and low-rank approximation method parameters. More...
	HODLR_Matrix::HODLR_Matrix (Eigen::SparseMatrix< double > &inputMatrix, int inputSizeThreshold, user_IndexTree &input_IndexTree, std::string LR_Method="PS_Sparse")
	This constructor initializes the class with a sparse matrix and a user defined indexing scheme used to create the the HODLR indexing tree. More...
	HODLR_Matrix::HODLR_Matrix (Eigen::SparseMatrix< double > &inputMatrix, Eigen::SparseMatrix< double > &inputGraph, int inputSizeThreshold, user_IndexTree &input_IndexTree, std::string LR_Method="PS_Sparse")
	This constructor initializes the class with a sparse matrix and an interaction graph (sparse matrix) and a user defined indexing schemes. The grpah is going to be used primarily as the interaction graph in the BDLR scheme. The user defined indexing scheme will be used to create the HODLR index tree. More...

Detailed Description

Thse constructors build an HODLR matrix from a preallocated sparse matrix.

Note

Currently, the sparse matrix is being passed by a non const reference variable. So if you want your original matrix, copy it elsewhere befor passing it to the constructor.

Only use "PS_Sparse" or "identifyBoundary" as low-rank approximation methods when initializing the class with a sparse matrix.

Function Documentation

HODLR_Matrix::HODLR_Matrix	(	Eigen::SparseMatrix< double > &	inputMatrix,
		int	inputSizeThreshold = 30,
		std::string	LR_Method = "PS_Sparse"
	)

This constructor initializes the class with a sparse matrix and optional leaf size and low-rank approximation method parameters.

Parameters

[in]	inputMatrix	: Preallocated Sparse Matrix.
[in]	inputSizeThreshold	: Leaf size threhold. If no value is provided, it will set the leaf size to the default value of 30.
[in]	LR_Method	: Low-rank approximation scheme to be used in calculating the off-diagonal low-rank approximations. If no value is provided, it will set the LR_Method parameter to "PS_Sparse".

HODLR_Matrix::HODLR_Matrix	(	Eigen::SparseMatrix< double > &	inputMatrix,
		int	inputSizeThreshold,
		user_IndexTree &	input_IndexTree,
		std::string	LR_Method = "PS_Sparse"
	)

This constructor initializes the class with a sparse matrix and a user defined indexing scheme used to create the the HODLR indexing tree.

Parameters

[in]	inputMatrix	: Preallocated Sparse Matrix.
[in]	inputSizeThreshold	: Leaf size threhold. If no value is provided, it will set the leaf size to the default value of 30.
[in]	input_IndexTree	: Leaf size threshold.
[in]	LR_Method	: Low-rank approximation scheme to be used in calculating the off-diagonal low-rank approximations. If no value is provided, it will set the LR_Method parameter to "PS_Sparse".

HODLR_Matrix::HODLR_Matrix	(	Eigen::SparseMatrix< double > &	inputMatrix,
		Eigen::SparseMatrix< double > &	inputGraph,
		int	inputSizeThreshold,
		user_IndexTree &	input_IndexTree,
		std::string	LR_Method = "PS_Sparse"
	)

This constructor initializes the class with a sparse matrix and an interaction graph (sparse matrix) and a user defined indexing schemes. The grpah is going to be used primarily as the interaction graph in the BDLR scheme. The user defined indexing scheme will be used to create the HODLR index tree.

Parameters

[in]	inputMatrix	: Preallocated Sparse Matrix.
[in]	inputGraph	: Preallocated sparse matrix. This sparse matrix will be used as the interaction graph in the BDLR low-rank approximation scheme.
[in]	inputSizeThreshold	: Leaf size threhold. If no value is provided, it will set the leaf size to the default value of 30.
[in]	input_IndexTree	: Leaf size threshold.
[in]	LR_Method	: Low-rank approximation scheme to be used in calculating the off-diagonal low-rank approximations. If no value is provided, it will set the LR_Method parameter to "PS_Sparse".