Compressed sparse row implementation of a sparse matrix. More...

#include <Kokkos_CrsMatrix.hpp>

Public Types
typedef CrsMatrix< ScalarType, OrdinalType, host_mirror_space, MemoryTraits >	HostMirror
	Type of a host-memory mirror of the sparse matrix. More...

typedef Kokkos::StaticCrsGraph < OrdinalType, Kokkos::LayoutLeft, Device, SizeType >	StaticCrsGraphType
	Type of the graph structure of the sparse matrix. More...

typedef StaticCrsGraphType::entries_type	index_type
	Type of column indices in the sparse matrix. More...

typedef StaticCrsGraphType::row_map_type	row_map_type
	Type of the "row map" (which contains the offset for each row's data). More...

typedef Kokkos::View < value_type *, Kokkos::LayoutRight, device_type, MemoryTraits >	values_type
	Kokkos Array type of the entries (values) in the sparse matrix. More...

typedef values_type::const_value_type	const_value_type
	Const version of the type of the entries in the sparse matrix. More...

typedef values_type::non_const_value_type	non_const_value_type
	Nonconst version of the type of the entries in the sparse matrix. More...

Public Member Functions
	CrsMatrix ()
	Default constructor; constructs an empty sparse matrix. More...

template<typename SType , typename OType , class DType , class MTType , typename IType >
	CrsMatrix (const CrsMatrix< SType, OType, DType, MTType, IType > &B)
	Copy Constructor. More...

	CrsMatrix (const std::string &arg_label, const StaticCrsGraphType &arg_graph)
	Construct with a graph that will be shared. More...

	CrsMatrix (const std::string &label, OrdinalType nrows, OrdinalType ncols, OrdinalType annz, ScalarType val, OrdinalType rows, OrdinalType *cols, bool pad=false)
	Constructor that copies raw arrays of host data in coordinate format. More...

	CrsMatrix (const std::string &label, OrdinalType nrows, OrdinalType ncols, OrdinalType annz, values_type vals, row_map_type rows, index_type cols)
	Constructor that accepts a row map, column indices, and values. More...

	CrsMatrix (const std::string &label, OrdinalType ncols, values_type vals, StaticCrsGraphType graph_)
	Constructor that accepts a a static graph, and values. More...

void	generate (const std::string &label, OrdinalType nrows, OrdinalType ncols, OrdinalType target_nnz, OrdinalType varianz_nel_row, OrdinalType width_row)
	This is a method only for testing that creates a random sparse matrix. More...

KOKKOS_INLINE_FUNCTION ordinal_type	numRows () const
	The number of rows in the sparse matrix. More...

KOKKOS_INLINE_FUNCTION ordinal_type	numCols () const
	The number of columns in the sparse matrix. More...

KOKKOS_INLINE_FUNCTION ordinal_type	nnz () const
	The number of stored entries in the sparse matrix. More...

KOKKOS_INLINE_FUNCTION SparseRowView< CrsMatrix >	row (int i) const
	Return a view of row i of the matrix. More...

KOKKOS_INLINE_FUNCTION SparseRowViewConst< CrsMatrix >	rowConst (int i) const
	Return a const view of row i of the matrix. More...

Detailed Description

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>
class Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >

Compressed sparse row implementation of a sparse matrix.

Template Parameters

ScalarType	The type of entries in the sparse matrix.
OrdinalType	The type of column indices in the sparse matrix.
Device	The Kokkos Device type.
MemoryTraits	Traits describing how Kokkos manages and accesses data. The default parameter suffices for most users.

"Crs" stands for "compressed row sparse." This is the phrase Trilinos traditionally uses to describe compressed sparse row storage for sparse matrices, as described, for example, in Saad (2nd ed.).

Definition at line 310 of file Kokkos_CrsMatrix.hpp.

Member Typedef Documentation

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef CrsMatrix<ScalarType, OrdinalType, host_mirror_space, MemoryTraits> Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::HostMirror

Type of a host-memory mirror of the sparse matrix.

Definition at line 321 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef Kokkos::StaticCrsGraph<OrdinalType, Kokkos::LayoutLeft, Device,SizeType> Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::StaticCrsGraphType

Type of the graph structure of the sparse matrix.

FIXME (mfh 29 Sep 2013) It doesn't make much sense to use int (the fourth template parameter of CrsArray below) as SizeType, if OrdinalType is bigger than int. We should use Kokkos::Impl::if_c to pick the default SizeType, possibly as follows:

typedef Impl::if_c< (sizeof(OrdinalType) >= sizeof(typename ViewTraits<OrdinalType*, Kokkos::LayoutLeft, Device, void>::size_type)),
  OrdinalType,
  typename ViewTraits<OrdinalType*, Kokkos::LayoutLeft, Device, void>::size_type >
size_type;

The first argument of if_c is a bool condition. If true, OrdinalType is size_type, else CrsArray's default size_type is size_type. We took the ViewTraits expression from the default value of the fourth template parameter of CrsArray. I have tested that the above expression compiles.

There is also some argument to be made that size_type should be chosen dynamically, as a function of the number of entries. It's entirely possible that a (very large) local sparse matrix could have dimensions (and therefore column indices) that fit in int32_t, but more entries than can be addressed by int32_t or even uint32_t.

(CRT 1 Oct 2013) approached the issue above by giving the matrix a fifth template parameter which propagates to the CrsArray (now StaticCrsGraph). This defaults to size_t. We still might look for a better solution though. Type of the graph structure of the sparse matrix.

Definition at line 357 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef StaticCrsGraphType::entries_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::index_type

Type of column indices in the sparse matrix.

Definition at line 360 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef StaticCrsGraphType::row_map_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::row_map_type

Type of the "row map" (which contains the offset for each row's data).

Definition at line 362 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef Kokkos::View<value_type*, Kokkos::LayoutRight, device_type, MemoryTraits> Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::values_type

Kokkos Array type of the entries (values) in the sparse matrix.

Definition at line 364 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef values_type::const_value_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::const_value_type

Const version of the type of the entries in the sparse matrix.

Definition at line 366 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

typedef values_type::non_const_value_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::non_const_value_type

Nonconst version of the type of the entries in the sparse matrix.

Definition at line 368 of file Kokkos_CrsMatrix.hpp.

Constructor & Destructor Documentation

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix ( )

inline

Default constructor; constructs an empty sparse matrix.

FIXME (mfh 09 Aug 2013) numRows, numCols, and nnz should be properties of the graph, not the matrix. Then CrsMatrix needs methods to get these from the graph.

Definition at line 388 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

template<typename SType , typename OType , class DType , class MTType , typename IType >

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix ( const CrsMatrix< SType, OType, DType, MTType, IType > & B )

inline

Copy Constructor.

Definition at line 400 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix	(	const std::string &	arg_label,
		const StaticCrsGraphType &	arg_graph
	)

inline

Construct with a graph that will be shared.

Allocate the values array for subsquent fill.

Definition at line 412 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix	(	const std::string &	label,
		OrdinalType	nrows,
		OrdinalType	ncols,
		OrdinalType	annz,
		ScalarType *	val,
		OrdinalType *	rows,
		OrdinalType *	cols,
		bool	pad = `false`
	)

inline

Constructor that copies raw arrays of host data in coordinate format.

On input, each entry of the sparse matrix is stored in val[k], with row index rows[k] and column index cols[k]. We assume that the entries are sorted in increasing order by row index.

This constructor is mainly useful for benchmarking or for reading the sparse matrix's data from a file.

Parameters

label	[in] The sparse matrix's label.
nrows	[in] The number of rows.
ncols	[in] The number of columns.
annz	[in] The number of entries.
val	[in] The entries.
rows	[in] The row indices. rows[k] is the row index of val[k].
cols	[in] The column indices. cols[k] is the column index of val[k].
pad	[in] If true, pad the sparse matrix's storage with zeros in order to improve cache alignment and / or vectorization.

FIXME (mfh 21 Jun 2013) The pad argument is currently not used.

Definition at line 447 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix	(	const std::string &	label,
		OrdinalType	nrows,
		OrdinalType	ncols,
		OrdinalType	annz,
		values_type	vals,
		row_map_type	rows,
		index_type	cols
	)

inline

Constructor that accepts a row map, column indices, and values.

The matrix will store and use the row map, indices, and values directly (by view, not by deep copy).

Parameters

label	[in] The sparse matrix's label.
nrows	[in] The number of rows.
ncols	[in] The number of columns.
annz	[in] The number of entries.
vals	[in/out] The entries.
rows	[in/out] The row map (containing the offsets to the data in each row).
cols	[in/out] The column indices.

Definition at line 486 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::CrsMatrix	(	const std::string &	label,
		OrdinalType	ncols,
		values_type	vals,
		StaticCrsGraphType	graph_
	)

inline

Constructor that accepts a a static graph, and values.

The matrix will store and use the row map, indices, and values directly (by view, not by deep copy).

Parameters

label	[in] The sparse matrix's label.
nrows	[in] The number of rows.
ncols	[in] The number of columns.
annz	[in] The number of entries.
vals	[in/out] The entries.
rows	[in/out] The row map (containing the offsets to the data in each row).
cols	[in/out] The column indices.

Definition at line 519 of file Kokkos_CrsMatrix.hpp.

Member Function Documentation

template<typename ScalarType , typename OrdinalType , class Device , class MemoryTraits , typename SizeType >

void Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::generate	(	const std::string &	label,
		OrdinalType	nrows,
		OrdinalType	ncols,
		OrdinalType	target_nnz,
		OrdinalType	varianz_nel_row,
		OrdinalType	width_row
	)

This is a method only for testing that creates a random sparse matrix.

Definition at line 805 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

KOKKOS_INLINE_FUNCTION ordinal_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::numRows ( ) const

inline

The number of rows in the sparse matrix.

Definition at line 657 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

KOKKOS_INLINE_FUNCTION ordinal_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::numCols ( ) const

inline

The number of columns in the sparse matrix.

Definition at line 662 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

KOKKOS_INLINE_FUNCTION ordinal_type Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::nnz ( ) const

inline

The number of stored entries in the sparse matrix.

Definition at line 667 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

KOKKOS_INLINE_FUNCTION SparseRowView<CrsMatrix> Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::row ( int i ) const

inline

Return a view of row i of the matrix.

Definition at line 675 of file Kokkos_CrsMatrix.hpp.

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>

KOKKOS_INLINE_FUNCTION SparseRowViewConst<CrsMatrix> Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >::rowConst ( int i ) const

inline

Return a const view of row i of the matrix.

Definition at line 693 of file Kokkos_CrsMatrix.hpp.

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

Kokkos_CrsMatrix.hpp

Public Types

Public Member Functions

Detailed Description

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t> class Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >

Member Typedef Documentation

Constructor & Destructor Documentation

Member Function Documentation

template<typename ScalarType, typename OrdinalType, class Device, class MemoryTraits = void, typename SizeType = size_t>
class Kokkos::CrsMatrix< ScalarType, OrdinalType, Device, MemoryTraits, SizeType >