Tpetra::Vector

Tpetra::Vector< Ordinal, Scalar > Class Template Reference

Tpetra::Vector: A class for constructing and using sparse vectors. More...

#include <Tpetra_VectorDecl.hpp>

Inheritance diagram for Tpetra::Vector< Ordinal, Scalar >:


Public Member Functions
Constructor/Destructor Methods
	Vector (const Map< Ordinal > &map)
	Sets all vector entries to zero.
	Vector (const Teuchos::ArrayView< const Scalar > &values, const Map< Ordinal > &map)
	Set object values from user array. Throws an exception if an incorrect number of entries are specified.
	Vector (const Vector< Ordinal, Scalar > &source)
	Copy constructor.
	~Vector ()
	Destructor.
Post-Construction Modification Routines
void	submitEntries (const Teuchos::ArrayView< const Ordinal > &indices, const Teuchos::ArrayView< const Scalar > &values)
	Submit entries. Values submitted will be summed with existing values.
void	setAllToScalar (const Scalar &value)
	Set all entries to scalarValue.
void	setAllToRandom ()
	Set all entries to random values.
Mathematical Methods
Scalar	dotProduct (const Vector< Ordinal, Scalar > &x) const
	Returns result of dot product, result = this.x.
void	absoluteValue (const Vector< Ordinal, Scalar > &x)
	Changes this vector to elementwise absolute values of x.
void	reciprocal (const Vector< Ordinal, Scalar > &x)
	Changes this vector to element-wise reciprocal values of x.
void	scale (const Scalar &scalarThis)
	Scale the current values of a vector, this = scalarThis*this.
void	scale (const Scalar &scalarX, const Vector< Ordinal, Scalar > &x)
	Replace vector values with scaled values of x, this = scalarX*x.
void	update (const Scalar &scalarX, const Vector< Ordinal, Scalar > &x, const Scalar &scalarThis)
	Update vector values with scaled values of x, this = scalarThis*this + scalarX*x.
void	update (const Scalar &scalarX, const Vector< Ordinal, Scalar > &x, const Scalar &scalarY, const Vector< Ordinal, Scalar > &y, const Scalar &scalarThis)
	Update vector with scaled values of x and y, this = scalarThis*this + scalarXx + scalarYy.
Scalar	norm1 () const
	Compute 1-norm of vector.
Scalar	norm2 () const
	Compute 2-norm of vector.
Scalar	normInf () const
	Compute Infinity-norm of vector.
Scalar	normWeighted (const Vector< Ordinal, Scalar > &weights) const
	Compute Weighted 2-norm (RMS Norm) of vector.
Scalar	minValue () const
	Compute minimum value of vector.
Scalar	maxValue () const
	Compute maximum value of vector.
Scalar	meanValue () const
	Compute mean (average) value of vector.
void	elementwiseMultiply (const Scalar &scalarXY, const Vector< Ordinal, Scalar > &x, const Vector< Ordinal, Scalar > &y, const Scalar &scalarThis)
	Vector multiplication (elementwise).
void	elementwiseReciprocalMultiply (Scalar scalarXY, const Vector< Ordinal, Scalar > &x, const Vector< Ordinal, Scalar > &y, const Scalar &scalarThis)
	Reciprocal multiply (elementwise).
Random number utilities
const Scalar &	getSeed () const
	Get seed.
void	setSeed (const Scalar &seed)
	Set seed.
Element access methods
Scalar &	operator[] (Ordinal index)
	[] operator, nonconst version
const Scalar &	operator[] (Ordinal index) const
	[] operator, const version
Attribute access methods
Ordinal	getNumMyEntries () const
	Returns number of vector entries owned by this image.
Ordinal	getNumGlobalEntries () const
	Returns number of vector entries across all images.
I/O methods
void	print (std::ostream &os) const
	Print method, used by overloaded << operator.
Misc.
const Map< Ordinal > &	getMap () const
	Returns a const reference to the VectorSpace this Vector belongs to.
Vector< Ordinal, Scalar > &	operator= (const Vector< Ordinal, Scalar > &source)
	Assignment Operator.
Expert/Developer Use Only.

Detailed Description

template<typename Ordinal, typename Scalar>
class Tpetra::Vector< Ordinal, Scalar >

Tpetra::Vector: A class for constructing and using sparse vectors.

Vector is templated on Scalar for the vector entries, and on Ordinal for the vector indices. A VectorSpace object is needed for all Vector objects.

Vector entries can only be accessed through their local index values. Global index values can be converted to local indices by using the VectorSpace::getLocalIndex method.

Note that for most of the mathematical methods that set this to the result of an operation on vectors passed as parameters, the this vector can be used as one of the parameters (unless otherwise specified).

Vector error codes (positive for non-fatal, negative for fatal):

+1 Specified vector index not found on this image.
+2 Vector sizes do not match.
+3 Cannot perform that operation on an empty vector.
-1 Invalid number of entries passed to constructor.
-99 Internal Vector error. Contact developer.