Template Numerical Library: TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView

Searching...

No Matches

RowView is a simple structure for accessing rows of sparse matrix. More...

#include <TNL/Matrices/SparseMatrixRowView.h>

Public Types
using	ColumnsIndexesViewType = ColumnsIndexesView
	Type of container view used for storing the column indexes of the matrix elements.
using	ConstColumnsIndexesViewType = typename ColumnsIndexesViewType::ConstViewType
	Type of constant container view used for storing the column indexes of the matrix elements.
using	ConstIteratorType = MatrixRowViewIterator< ConstRowView >
	Type of constant iterator for the matrix row.
using	ConstRowView = SparseMatrixRowView< SegmentViewType, ConstValuesViewType, ConstColumnsIndexesViewType >
	Type of constant sparse matrix row view.
using	ConstValuesViewType = typename ValuesViewType::ConstViewType
	Type of constant container view used for storing the matrix elements values.
using	GetValueConstResultType = std::conditional_t< isBinary(), bool, std::add_const_t< RealType >& >
using	GetValueResultType = std::conditional_t< isBinary(), bool, RealType& >
using	IndexType = typename ColumnsIndexesView::IndexType
	The type used for matrix elements indexing.
using	IteratorType = MatrixRowViewIterator< RowView >
	Type of iterator for the matrix row.
using	MatrixElementType = SparseMatrixElement< RealType, IndexType >
	The type of related matrix element.
using	RealType = typename ValuesView::RealType
	The type of matrix elements.
using	RowView = SparseMatrixRowView< SegmentViewType, ValuesViewType, ColumnsIndexesViewType >
	Type of sparse matrix row view.
using	SegmentViewType = SegmentView
	Type representing matrix row format.
using	ValuesViewType = ValuesView
	Type of container view used for storing the matrix elements values.

Public Member Functions
__cuda_callable__	SparseMatrixRowView (SegmentViewType segmentView, ValuesViewType values, ColumnsIndexesViewType columnIndexes)
	Constructor with segmentView, values and columnIndexes.
__cuda_callable__ IteratorType	begin ()
	Returns iterator pointing at the beginning of the matrix row.
__cuda_callable__ ConstIteratorType	cbegin () const
	Returns constant iterator pointing at the beginning of the matrix row.
__cuda_callable__ ConstIteratorType	cend () const
	Returns constant iterator pointing at the end of the matrix row.
__cuda_callable__ IteratorType	end ()
	Returns iterator pointing at the end of the matrix row.
__cuda_callable__ ColumnsIndexesViewType::ValueType &	getColumnIndex (IndexType localIdx)
	Returns non-constants reference to a column index of an element with given rank in the row.
__cuda_callable__ const ColumnsIndexesViewType::ValueType &	getColumnIndex (IndexType localIdx) const
	Returns constants reference to a column index of an element with given rank in the row.
__cuda_callable__ IndexType	getGlobalIndex (IndexType localIdx) const
	Computes the global index of the matrix element with given rank in the matrix row.
__cuda_callable__ IndexType	getRowIndex () const
	Returns the matrix row index.
__cuda_callable__ IndexType	getSize () const
	Returns size of the matrix row, i.e. number of matrix elements in this row.
__cuda_callable__ auto	getValue (IndexType localIdx) -> GetValueResultType
	Returns non-constants reference to value of an element with given rank in the row.
__cuda_callable__ auto	getValue (IndexType localIdx) const -> GetValueConstResultType
	Returns constants reference to value of an element with given rank in the row.
template<typename _SegmentView, typename _ValuesView, typename _ColumnsIndexesView>
__cuda_callable__ bool	operator== (const SparseMatrixRowView< _SegmentView, _ValuesView, _ColumnsIndexesView > &other) const
	Comparison of two matrix rows.
__cuda_callable__ void	setColumnIndex (IndexType localIdx, const IndexType &columnIndex)
	Sets a column index of matrix element with given rank in the matrix row.
__cuda_callable__ void	setElement (IndexType localIdx, IndexType columnIndex, const RealType &value)
	Sets both a value and a column index of matrix element with given rank in the matrix row.
__cuda_callable__ void	setValue (IndexType localIdx, const RealType &value)
	Sets a value of matrix element with given rank in the matrix row.
__cuda_callable__ void	sortColumnIndexes ()
	Sort the matrix row by column indexes in ascending order.

Static Public Member Functions
static constexpr bool	isBinary ()
	Tells whether the parent matrix is a binary matrix.

Protected Attributes
ColumnsIndexesViewType	columnIndexes
SegmentViewType	segmentView
ValuesViewType	values

Detailed Description

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>
class TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >

RowView is a simple structure for accessing rows of sparse matrix.

Template Parameters

SegmentView	is a segment view of segments representing the matrix format.
ValuesView	is a vector view storing the matrix elements values.
ColumnsIndexesView	is a vector view storing the column indexes of the matrix element.

See SparseMatrixBase, SparseMatrix and SparseMatrixView.

Example: #include <iostream>

#include <TNL/Algorithms/parallelFor.h>

#include <TNL/Matrices/SparseMatrix.h>

#include <TNL/Devices/Host.h>

#include <TNL/Devices/Cuda.h>

template< typename Device >

void

getRowExample()

{

/***

* Set the following matrix (dots represent zero matrix elements):

*

* / 2 . . . . \

* | 1 2 1 . . |

* | . 1 2 1. . |

* | . . 1 2 1 |

* \ . . . . 2 /

*/

const int size = 5;

TNL::Matrices::SparseMatrix< double, Device > matrix( { 1, 3, 3, 3, 1 }, size );

auto view = matrix.getView();

/***

* Set the matrix elements.

*/

auto f = [ = ] __cuda_callable__( int rowIdx ) mutable

{

auto row = view.getRow( rowIdx );

if( rowIdx == 0 ) {

row.setElement( 0, rowIdx, 2.0 ); // diagonal element

}

else if( rowIdx == size - 1 ) {

row.setElement( 0, rowIdx, 2.0 ); // diagonal element

}

else {

row.setElement( 0, rowIdx - 1, 1.0 ); // elements below the diagonal

row.setElement( 1, rowIdx, 2.0 ); // diagonal element

row.setElement( 2, rowIdx + 1, 1.0 ); // elements above the diagonal

}

};

TNL::Algorithms::parallelFor< Device >( 0, matrix.getRows(), f );

std::cout << matrix << '\n';

}

int

main( int argc, char* argv[] )

{

std::cout << "Getting matrix rows on host:\n";

getRowExample< TNL::Devices::Host >();

#ifdef __CUDACC__

std::cout << "Getting matrix rows on CUDA device:\n";

getRowExample< TNL::Devices::Cuda >();

#endif

}

__cuda_callable__
#define __cuda_callable__
This macro serves for annotating functions which are supposed to be called even from the GPU device.
Definition Macros.h:50

std::cout

TNL::Matrices::SparseMatrixBase::getRow
__cuda_callable__ ConstRowView getRow(IndexType rowIdx) const
Constant getter of simple structure for accessing given matrix row.

TNL::Matrices::SparseMatrix
Implementation of sparse matrix, i.e. matrix storing only non-zero elements.
Definition SparseMatrix.h:59

TNL::Matrices::SparseMatrix::getView
ViewType getView()
Returns a modifiable view of the sparse matrix.

TNL::Algorithms::parallelFor
std::enable_if_t< std::is_integral_v< Begin > &&std::is_integral_v< End > > parallelFor(const Begin &begin, const End &end, typename Device::LaunchConfiguration launch_config, Function f, FunctionArgs... args)
Parallel for-loop function for 1D range specified with integral values.
Definition parallelFor.h:41

Output: Getting matrix rows on host:

Row: 0 -> 0:2

Row: 1 -> 0:1 1:2 2:1

Row: 2 -> 1:1 2:2 3:1

Row: 3 -> 2:1 3:2 4:1

Row: 4 -> 4:2

Getting matrix rows on CUDA device:

Row: 0 -> 0:2

Row: 1 -> 0:1 1:2 2:1

Row: 2 -> 1:1 2:2 3:1

Row: 3 -> 2:1 3:2 4:1

Row: 4 -> 4:2

Constructor & Destructor Documentation

◆ SparseMatrixRowView()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::SparseMatrixRowView	(	SegmentViewType	segmentView,
		ValuesViewType	values,
		ColumnsIndexesViewType	columnIndexes )

Constructor with segmentView, values and columnIndexes.

Parameters

segmentView	instance of SegmentViewType representing matrix row.
values	is a container view for storing the matrix elements values.
columnIndexes	is a container view for storing the column indexes of the matrix elements.

Member Function Documentation

◆ begin()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ IteratorType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::begin ( )

nodiscard

Returns iterator pointing at the beginning of the matrix row.

Returns: iterator pointing at the beginning.

◆ cbegin()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ ConstIteratorType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::cbegin ( ) const

nodiscard

Returns constant iterator pointing at the beginning of the matrix row.

Returns: iterator pointing at the beginning.

◆ cend()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ ConstIteratorType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::cend ( ) const

nodiscard

Returns constant iterator pointing at the end of the matrix row.

Returns: iterator pointing at the end.

◆ end()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ IteratorType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::end ( )

nodiscard

Returns iterator pointing at the end of the matrix row.

Returns: iterator pointing at the end.

◆ getColumnIndex() [1/2]

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ ColumnsIndexesViewType::ValueType & TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getColumnIndex ( IndexType localIdx )

nodiscard

Returns non-constants reference to a column index of an element with given rank in the row.

Parameters

localIdx is the rank of the non-zero element in given row.

Returns: non-constant reference to the matrix element column index.

◆ getColumnIndex() [2/2]

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ const ColumnsIndexesViewType::ValueType & TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getColumnIndex ( IndexType localIdx ) const

nodiscard

Returns constants reference to a column index of an element with given rank in the row.

Parameters

localIdx is the rank of the non-zero element in given row.

Returns: constant reference to the matrix element column index.

◆ getGlobalIndex()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ IndexType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getGlobalIndex ( IndexType localIdx ) const

nodiscard

Computes the global index of the matrix element with given rank in the matrix row.

Parameters

localIdx is the rank of the matrix element in the row.

Returns: global index of the matrix element.

◆ getRowIndex()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ IndexType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getRowIndex ( ) const

nodiscard

Returns the matrix row index.

Returns: matrix row index.

◆ getSize()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ IndexType TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getSize ( ) const

nodiscard

Returns size of the matrix row, i.e. number of matrix elements in this row.

Returns: Size of the matrix row.

◆ getValue() [1/2]

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ auto TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getValue ( IndexType localIdx ) -> GetValueResultType

nodiscard

Returns non-constants reference to value of an element with given rank in the row.

Parameters

localIdx is the rank of the non-zero element in given row.

Returns: non-constant reference to the matrix element value.

◆ getValue() [2/2]

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ auto TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::getValue ( IndexType localIdx ) const -> GetValueConstResultType

nodiscard

Returns constants reference to value of an element with given rank in the row.

Parameters

localIdx is the rank of the non-zero element in given row.

Returns: constant reference to the matrix element value.

◆ isBinary()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

constexpr bool TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::isBinary ( )

inlinestaticnodiscardconstexpr

Tells whether the parent matrix is a binary matrix.

Returns: true if the matrix is binary.

◆ operator==()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

template<typename _SegmentView, typename _ValuesView, typename _ColumnsIndexesView>

__cuda_callable__ bool TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::operator== ( const SparseMatrixRowView< _SegmentView, _ValuesView, _ColumnsIndexesView > & other ) const

nodiscard

Comparison of two matrix rows.

The other matrix row can be from any other matrix. Only elements which are really present in the matrix are compared, i.e. the padding elements and the elements with an explicitly stored zero value are skipped, so two rows are equal even if they differ in the explicitly stored zero elements only. The elements are compared in the order in which they are stored in the rows, i.e. the rows do not need to be sorted by the column indexes, but rows which store the same elements in a different order are not equal.

Parameters

other is another matrix row.

Returns: true if both rows are the same, false otherwise.

◆ setColumnIndex()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ void TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::setColumnIndex	(	IndexType	localIdx,
		const IndexType &	columnIndex )

Sets a column index of matrix element with given rank in the matrix row.

Parameters

localIdx	is the rank of the matrix element in the row.
columnIndex	is the new column index of the matrix element.

◆ setElement()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ void TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::setElement	(	IndexType	localIdx,
		IndexType	columnIndex,
		const RealType &	value )

Sets both a value and a column index of matrix element with given rank in the matrix row.

Parameters

localIdx	is the rank of the matrix element in the row.
columnIndex	is the new column index of the matrix element.
value	is the new value of the matrix element.

◆ setValue()

template<typename SegmentView, typename ValuesView, typename ColumnsIndexesView>

__cuda_callable__ void TNL::Matrices::SparseMatrixRowView< SegmentView, ValuesView, ColumnsIndexesView >::setValue	(	IndexType	localIdx,
		const RealType &	value )

Sets a value of matrix element with given rank in the matrix row.

Parameters

localIdx	is the rank of the matrix element in the row.
value	is the new value of the matrix element.

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

src/TNL/Matrices/SparseMatrixRowView.h

Public Types

Public Member Functions

Static Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ SparseMatrixRowView()

Member Function Documentation

◆ begin()

◆ cbegin()

◆ cend()

◆ end()

◆ getColumnIndex() [1/2]

◆ getColumnIndex() [2/2]

◆ getGlobalIndex()

◆ getRowIndex()

◆ getSize()

◆ getValue() [1/2]

◆ getValue() [2/2]

◆ isBinary()

◆ operator==()

◆ setColumnIndex()

◆ setElement()

◆ setValue()