ViennaCL - The Vienna Computing Library  1.4.2
Public Types | Public Member Functions
compressed_matrix< SCALARTYPE, ALIGNMENT > Class Template Reference

A sparse square matrix in compressed sparse rows format. More...

#include <compressed_matrix.hpp>

Public Types

typedef
viennacl::backend::mem_handle 
handle_type
typedef scalar< typename
viennacl::tools::CHECK_SCALAR_TEMPLATE_ARGUMENT
< SCALARTYPE >::ResultType > 
value_type
typedef vcl_size_t size_type

Public Member Functions

 compressed_matrix ()
 Default construction of a compressed matrix. No memory is allocated.
 compressed_matrix (std::size_t rows, std::size_t cols, std::size_t nonzeros=0)
 Construction of a compressed matrix with the supplied number of rows and columns. If the number of nonzeros is positive, memory is allocated.
compressed_matrixoperator= (compressed_matrix const &other)
 Assignment a compressed matrix from possibly another memory domain.
void set (const void *row_jumper, const void *col_buffer, const SCALARTYPE *elements, std::size_t rows, std::size_t cols, std::size_t nonzeros)
 Sets the row, column and value arrays of the compressed matrix.
void reserve (std::size_t new_nonzeros)
 Allocate memory for the supplied number of nonzeros in the matrix. Old values are preserved.
void resize (std::size_t new_size1, std::size_t new_size2, bool preserve=true)
 Resize the matrix.
entry_proxy< SCALARTYPE > operator() (std::size_t i, std::size_t j)
 Returns a reference to the (i,j)-th entry of the sparse matrix. If (i,j) does not exist (zero), it is inserted (slow!)
const std::size_t & size1 () const
 Returns the number of rows.
const std::size_t & size2 () const
 Returns the number of columns.
const std::size_t & nnz () const
 Returns the number of nonzero entries.
const handle_typehandle1 () const
 Returns the OpenCL handle to the row index array.
const handle_typehandle2 () const
 Returns the OpenCL handle to the column index array.
const handle_typehandle () const
 Returns the OpenCL handle to the matrix entry array.
handle_typehandle1 ()
 Returns the OpenCL handle to the row index array.
handle_typehandle2 ()
 Returns the OpenCL handle to the column index array.
handle_typehandle ()
 Returns the OpenCL handle to the matrix entry array.
void switch_memory_domain (viennacl::memory_types new_domain)
viennacl::memory_types memory_domain () const

Detailed Description

template<class SCALARTYPE, unsigned int ALIGNMENT>
class viennacl::compressed_matrix< SCALARTYPE, ALIGNMENT >

A sparse square matrix in compressed sparse rows format.

Template Parameters:
SCALARTYPEThe floating point type (either float or double, checked at compile time)
ALIGNMENTThe internal memory size for the entries in each row is given by (size()/ALIGNMENT + 1) * ALIGNMENT. ALIGNMENT must be a power of two. Best values or usually 4, 8 or 16, higher values are usually a waste of memory.

Member Typedef Documentation

typedef scalar<typename viennacl::tools::CHECK_SCALAR_TEMPLATE_ARGUMENT<SCALARTYPE>::ResultType> value_type

Constructor & Destructor Documentation

compressed_matrix ( ) [inline]

Default construction of a compressed matrix. No memory is allocated.

compressed_matrix ( std::size_t  rows,
std::size_t  cols,
std::size_t  nonzeros = 0 
) [inline, explicit]

Construction of a compressed matrix with the supplied number of rows and columns. If the number of nonzeros is positive, memory is allocated.

Parameters:
rowsNumber of rows
colsNumber of columns
nonzerosOptional number of nonzeros for memory preallocation

Member Function Documentation

const handle_type& handle ( ) const [inline]

Returns the OpenCL handle to the matrix entry array.

handle_type& handle ( ) [inline]

Returns the OpenCL handle to the matrix entry array.

const handle_type& handle1 ( ) const [inline]

Returns the OpenCL handle to the row index array.

handle_type& handle1 ( ) [inline]

Returns the OpenCL handle to the row index array.

const handle_type& handle2 ( ) const [inline]

Returns the OpenCL handle to the column index array.

handle_type& handle2 ( ) [inline]

Returns the OpenCL handle to the column index array.

const std::size_t& nnz ( ) const [inline]

Returns the number of nonzero entries.

entry_proxy<SCALARTYPE> operator() ( std::size_t  i,
std::size_t  j 
) [inline]

Returns a reference to the (i,j)-th entry of the sparse matrix. If (i,j) does not exist (zero), it is inserted (slow!)

compressed_matrix& operator= ( compressed_matrix< SCALARTYPE, ALIGNMENT > const &  other) [inline]

Assignment a compressed matrix from possibly another memory domain.

void reserve ( std::size_t  new_nonzeros) [inline]

Allocate memory for the supplied number of nonzeros in the matrix. Old values are preserved.

void resize ( std::size_t  new_size1,
std::size_t  new_size2,
bool  preserve = true 
) [inline]

Resize the matrix.

Parameters:
new_size1New number of rows
new_size2New number of columns
preserveIf true, the old values are preserved. At present, old values are always discarded.
void set ( const void *  row_jumper,
const void *  col_buffer,
const SCALARTYPE *  elements,
std::size_t  rows,
std::size_t  cols,
std::size_t  nonzeros 
) [inline]

Sets the row, column and value arrays of the compressed matrix.

Parameters:
row_jumperPointer to an array holding the indices of the first element of each row (starting with zero). E.g. row_jumper[10] returns the index of the first entry of the 11th row. The array length is 'cols + 1'
col_bufferPointer to an array holding the column index of each entry. The array length is 'nonzeros'
elementsPointer to an array holding the entries of the sparse matrix. The array length is 'elements'
rowsNumber of rows of the sparse matrix
colsNumber of columns of the sparse matrix
nonzerosNumber of nonzeros
const std::size_t& size1 ( ) const [inline]

Returns the number of rows.

const std::size_t& size2 ( ) const [inline]

Returns the number of columns.

void switch_memory_domain ( viennacl::memory_types  new_domain) [inline]

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