NOX Development
Public Member Functions | Protected Member Functions | Protected Attributes
LOCA::Epetra::LowRankUpdateRowMatrix Class Reference

An Epetra row matrix for implementing the operator $P = J + U V^T$. More...

#include <LOCA_Epetra_LowRankUpdateRowMatrix.H>

Inheritance diagram for LOCA::Epetra::LowRankUpdateRowMatrix:
Inheritance graph
[legend]
Collaboration diagram for LOCA::Epetra::LowRankUpdateRowMatrix:
Collaboration graph
[legend]

List of all members.

Public Member Functions

 LowRankUpdateRowMatrix (const Teuchos::RCP< LOCA::GlobalData > &global_data, const Teuchos::RCP< Epetra_RowMatrix > &jacRowMatrix, const Teuchos::RCP< Epetra_MultiVector > &U_multiVec, const Teuchos::RCP< Epetra_MultiVector > &V_multiVec, bool setup_for_solve, bool include_UV_terms)
 Constructor.
virtual ~LowRankUpdateRowMatrix ()
 Destructor.
virtual const Epetra_BlockMapMap () const
 Returns a reference to the Epetra_BlockMap for this object.
virtual int NumMyRowEntries (int MyRow, int &NumEntries) const
 Returns the number of nonzero entries in MyRow.
virtual int MaxNumEntries () const
 Returns the maximum of NumMyRowEntries() over all rows.
virtual int ExtractMyRowCopy (int MyRow, int Length, int &NumEntries, double *Values, int *Indices) const
 Returns a copy of the specified local row in user-provided arrays.
virtual int ExtractDiagonalCopy (Epetra_Vector &Diagonal) const
 Returns a copy of the main diagonal in a user-provided vector.
virtual int Multiply (bool TransA, const Epetra_MultiVector &X, Epetra_MultiVector &Y) const
 Returns the result of a Epetra_RowMatrix multiplied by a Epetra_MultiVector X in Y.
virtual int Solve (bool Upper, bool Trans, bool UnitDiagonal, const Epetra_MultiVector &X, Epetra_MultiVector &Y) const
 Returns result of a local-only solve using a triangular Epetra_RowMatrix with Epetra_MultiVectors X and Y.
virtual int InvRowSums (Epetra_Vector &x) const
 Computes the sum of absolute values of the rows of the Epetra_RowMatrix, results returned in x.
virtual int LeftScale (const Epetra_Vector &x)
 Scales the Epetra_RowMatrix on the left with a Epetra_Vector x.
virtual int InvColSums (Epetra_Vector &x) const
 Computes the sum of absolute values of the columns of the Epetra_RowMatrix, results returned in x.
virtual int RightScale (const Epetra_Vector &x)
 Scales the Epetra_RowMatrix on the right with a Epetra_Vector x.
virtual bool Filled () const
 If FillComplete() has been called, this query returns true, otherwise it returns false.
virtual double NormInf () const
 Returns the infinity norm of the global matrix.
virtual double NormOne () const
 Returns the one norm of the global matrix.
virtual int NumGlobalNonzeros () const
 Returns the number of nonzero entries in the global matrix.
virtual int NumGlobalRows () const
 Returns the number of global matrix rows.
virtual int NumGlobalCols () const
 Returns the number of global matrix columns.
virtual int NumGlobalDiagonals () const
 Returns the number of global nonzero diagonal entries, based on global row/column index comparisons.
virtual int NumMyNonzeros () const
 Returns the number of nonzero entries in the calling processor's portion of the matrix.
virtual int NumMyRows () const
 Returns the number of matrix rows owned by the calling processor.
virtual int NumMyCols () const
 Returns the number of matrix columns owned by the calling processor.
virtual int NumMyDiagonals () const
 Returns the number of local nonzero diagonal entries, based on global row/column index comparisons.
virtual bool LowerTriangular () const
 If matrix is lower triangular in local index space, this query returns true, otherwise it returns false.
virtual bool UpperTriangular () const
 If matrix is upper triangular in local index space, this query returns true, otherwise it returns false.
virtual const Epetra_MapRowMatrixRowMap () const
 Returns the Epetra_Map object associated with the rows of this matrix.
virtual const Epetra_MapRowMatrixColMap () const
 Returns the Epetra_Map object associated with the columns of this matrix.
virtual const Epetra_ImportRowMatrixImporter () const
 Returns the Epetra_Import object that contains the import operations for distributed operations.

Protected Member Functions

double computeUV (int MyRow, int MyCol) const
 Compute MyRow, MyCol entry of $U V^T$.

Protected Attributes

Teuchos::RCP< Epetra_RowMatrixJ_rowMatrix
 Stores row matrix representing J.
Teuchos::RCP< Epetra_MultiVectornonconst_U
 Stores pointer to non-const U.
Teuchos::RCP< Epetra_MultiVectornonconst_V
 Stores pointer to non-const V.
bool includeUV
 Flag indicating whether to include U*V^T terms.
int m
 Number of columns in U and V.
const Epetra_BlockMapU_map
 Map for U.
const Epetra_BlockMapV_map
 Map for V.
const Epetra_BlockMaprow_map
 Row map for J.

Detailed Description

An Epetra row matrix for implementing the operator $P = J + U V^T$.

This class implements the Epetra_RowMatrix interface for $P = J + U V^T$ where $J$ is an Epetra_RowMatrix and $U$ and $V$ are Epetra_MultiVectors. It is derived from LOCA::Epetra::LowRankUpdateOp to implement the Epetra_Operator interface. The interface here implements the Epetra_RowMatrix interface when the matrix $J$ is itself a row matrix. This allows preconditioners to be computed and scaling in linear systems to be performed when using this operator. The implementation here merely adds the corresponding entries for $U V^T$ to the rows of $J$. Note however this is only an approximation to the true matrix $J + U V^T$.

This class assumes $U$ and $V$ have the same distribution as the rows of $J$.


Constructor & Destructor Documentation

LOCA::Epetra::LowRankUpdateRowMatrix::LowRankUpdateRowMatrix ( const Teuchos::RCP< LOCA::GlobalData > &  global_data,
const Teuchos::RCP< Epetra_RowMatrix > &  jacRowMatrix,
const Teuchos::RCP< Epetra_MultiVector > &  U_multiVec,
const Teuchos::RCP< Epetra_MultiVector > &  V_multiVec,
bool  setup_for_solve,
bool  include_UV_terms 
)

Constructor.

Parameters:
global_data[in] The global data object
jacRowMatrix[in] Jacobian operator J as a row matrix
U_multiVec[in] Multivector representing U
V_multiVec[in] Multivector representing V
setup_for_solve[in] Setup data structures for ApplyInverse()
include_UV_terms[in] Include $U V^T$ terms in RowMatrix routines ExtractRowCopy(), ExtactDiagonalCopy(), InvRowSums(), InvColSums(), NormInf() and NormOne().

The documentation for this class was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends