Implementation of NOX::Abstract::Vector for Petsc vectors. More...

#include <NOX_Petsc_Vector.H>

Inheritance diagram for NOX::Petsc::Vector:

Collaboration diagram for NOX::Petsc::Vector:

Public Member Functions
	Vector (const Vec &source, CopyType type=DeepCopy)
	Construct by copying map and/or elements of a Petsc Vec.

	Vector (const Vec &source, std::string name="Unnamed", CopyType type=DeepCopy)
	Construct by copying map and/or elements of a Petsc Vec.

	Vector (const Vector &source, CopyType type=DeepCopy)
	Construct by copying a NOX::Petsc::Vector.

	~Vector ()
	Destruct Vector.

virtual NOX::size_type	length () const
	Return the length of vector. More...

std::string	Name ()


virtual Vec &	getPetscVector ()
	Get reference to underlying Petsc vector.

virtual const Vec &	getPetscVector () const
	Get const reference to underlying Petsc vector.


virtual Abstract::Vector &	init (double value)
	Initialize every element of this vector with `gamma`. More...

virtual Abstract::Vector &	operator= (const Vec &source)
	Copies source vector into "this".

virtual Abstract::Vector &	operator= (const Vector &source)

virtual Abstract::Vector &	operator= (const Abstract::Vector &source)
	See above.

virtual Abstract::Vector &	abs (const Vector &source)

virtual Abstract::Vector &	abs (const Abstract::Vector &source)
	See above.

virtual Abstract::Vector &	random (bool useSeed=false, int seed=1)
	Initialize each element of this vector with a random value. More...

virtual Abstract::Vector &	reciprocal (const Vector &source)

virtual Abstract::Vector &	reciprocal (const Abstract::Vector &source)
	See above.


virtual Abstract::Vector &	scale (double gammaval)
	Scale each element of this vector by `gamma`. More...

virtual Abstract::Vector &	scale (const Vector &a)

virtual Abstract::Vector &	scale (const Abstract::Vector &a)
	See above.

virtual Abstract::Vector &	update (double alpha, const Vector &a, double gammaval=0.0)

virtual Abstract::Vector &	update (double alpha, const Abstract::Vector &a, double gammaval=0.0)
	See above.

virtual Abstract::Vector &	update (double alpha, const Vector &a, double beta, const Vector &b, double gammaval=0.0)

virtual Abstract::Vector &	update (double alpha, const Abstract::Vector &a, double beta, const Abstract::Vector &b, double gammaval=0.0)
	See above.


virtual Teuchos::RCP < NOX::Abstract::Vector >	clone (CopyType type=DeepCopy) const
	Create a new Vector of the same underlying type by cloning "this", and return a pointer to the new vector. More...


virtual double	norm (Abstract::Vector::NormType type=TwoNorm) const
	Norm. More...

virtual double	norm (const Vector &weights) const

virtual double	norm (const Abstract::Vector &weights) const
	See above.


virtual double	innerProduct (const Vector &y) const

virtual double	innerProduct (const Abstract::Vector &y) const
	See above.

Public Member Functions inherited from NOX::Abstract::Vector
	Vector ()
	Abstract Vector constructor (does nothing)

virtual void	print (std::ostream &stream) const
	Print the vector. To be used for debugging only.

virtual Teuchos::RCP < NOX::Abstract::MultiVector >	createMultiVector (const NOX::Abstract::Vector const vecs, int numVecs, NOX::CopyType type=NOX::DeepCopy) const
	Create a MultiVector with `numVecs+1` columns out of an array of Vectors. The vector stored under `this` will be the first column with the remaining `numVecs` columns given by `vecs`. More...

virtual Teuchos::RCP < NOX::Abstract::MultiVector >	createMultiVector (int numVecs, NOX::CopyType type=NOX::DeepCopy) const
	Create a MultiVector with `numVecs` columns. More...

Vec	petscVec
	Pointer to Petsc vector owned by this object.

bool	isAlloc
	Flag to indicate if the vector has been allocated.

std::string	name

virtual int	allocate (const Vec &source, CopyType type=DeepCopy)

Additional Inherited Members
Public Types inherited from NOX::Abstract::Vector
enum	NormType { TwoNorm, OneNorm, MaxNorm }
	Norm types used in norm() calculations. More...

Detailed Description

Implementation of NOX::Abstract::Vector for Petsc vectors.

Member Function Documentation

Teuchos::RCP< NOX::Abstract::Vector > Vector::clone ( CopyType type = DeepCopy ) const

virtual

Create a new Vector of the same underlying type by cloning "this", and return a pointer to the new vector.

If type is NOX::DeepCopy, then we need to create an exact replica of "this". Otherwise, if type is NOX::ShapeCopy, we need only replicate the shape of "this" (the memory is allocated for the objects, but the current values are not copied into the vector). Note that there is no assumption that a vector created by ShapeCopy is initialized to zeros.

Returns: Pointer to newly created vector or NULL if clone is not supported.

Implements NOX::Abstract::Vector.

References petscVec, and Teuchos::rcp().

Abstract::Vector & Vector::init ( double gamma )

virtual

Initialize every element of this vector with gamma.

Here x represents this vector, and we update it as

$x_i = \gamma \quad \mbox{for } i=1,\dots,n$

Returns: Reference to this object

Implements NOX::Abstract::Vector.

References petscVec.

NOX::size_type Vector::length ( ) const

virtual

Return the length of vector.

Returns: The length of this vector

Note: Even if the vector is distributed across processors, this should return the global length of the vector.

Implements NOX::Abstract::Vector.

References petscVec.

double Vector::norm ( Abstract::Vector::NormType type = TwoNorm ) const

virtual

Norm.

Here x represents this vector, and we compute its norm as follows: for each NOX::Abstract::Vector::NormType:

NOX::Abstract::Vector::TwoNorm
$\|x\| = \sqrt{\sum_{i=1}^{n} x_i^2}$
NOX::Abstract::Vector::OneNorm
$\|x\| = \sum_{i=1}^{n} |x_i|$
NOX::Abstract::Vector::MaxNorm
$\|x\| = \max_{i} |x_i|$

Returns: $\|x\|$

Implements NOX::Abstract::Vector.

References NOX::Abstract::Vector::MaxNorm, NOX::Abstract::Vector::OneNorm, petscVec, and NOX::Abstract::Vector::TwoNorm.

Referenced by NOX::Petsc::Group::computeF(), and norm().

Abstract::Vector & Vector::random	(	bool	useSeed = `false`,
		int	seed = `1`
	)

virtual

Initialize each element of this vector with a random value.

If useSeed is true, uses the value of seed to seed the random number generator before filling the entries of this vector. So, if two calls are made where useSeed is true and seed is the same, then the vectors returned should be the same.

Default implementation throw an error. Only referenced by LOCA methods.