Amesos  Development
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Groups Pages
Protected Member Functions | Protected Attributes | List of all members
Amesos_Lapack Class Reference

Amesos_Lapack: an interface to LAPACK. More...

#include <Amesos_Lapack.h>

Inheritance diagram for Amesos_Lapack:
Inheritance graph
[legend]
Collaboration diagram for Amesos_Lapack:
Collaboration graph
[legend]

Public Member Functions

 Amesos_Lapack (const Epetra_LinearProblem &LinearProblem)
 Amesos_Lapack Constructor. More...
 
 ~Amesos_Lapack (void)
 Amesos_Lapack Destructor. More...
 
int SymbolicFactorization ()
 Performs SymbolicFactorization on the matrix A. More...
 
int NumericFactorization ()
 Performs NumericFactorization on the matrix A. More...
 
int Solve ()
 Solves A X = B (or AT x = B) More...
 
const Epetra_LinearProblemGetProblem () const
 Returns the Epetra_LinearProblem. More...
 
bool MatrixShapeOK () const
 Returns true if the solver can handle this matrix shape. More...
 
int SetUseTranspose (bool UseTranspose_in)
 If set true, X will be set to the solution of AT X = B (not A X = B) More...
 
bool UseTranspose () const
 Returns the current UseTranspose setting.
 
const Epetra_CommComm () const
 Returns a pointer to the Epetra_Comm communicator associated with this operator.
 
void setParameterList (Teuchos::RCP< Teuchos::ParameterList > const &paramList)
 Use this parameter list to read values from. More...
 
Teuchos::RCP
< Teuchos::ParameterList
unsetParameterList ()
 This is an empty stub.
 
int SetParameters (Teuchos::ParameterList &ParameterList)
 Deprecated - Sets parameters.
 
int GEEV (Epetra_Vector &Er, Epetra_Vector &Ei)
 Computes the eigenvalues of the linear system matrix using DGEEV. More...
 
int NumSymbolicFact () const
 Returns the number of symbolic factorizations performed by this object.
 
int NumNumericFact () const
 Returns the number of numeric factorizations performed by this object.
 
int NumSolve () const
 Returns the number of solves performed by this object.
 
void PrintTiming () const
 Print timing information.
 
void PrintStatus () const
 Print information about the factorization and solution phases.
 
void GetTiming (Teuchos::ParameterList &TimingParameterList) const
 Extracts timing information from the current solver and places it in the parameter list.
 
- Public Member Functions inherited from Amesos_BaseSolver
virtual ~Amesos_BaseSolver ()
 Destructor.
 
virtual Teuchos::RCP
< Teuchos::ParameterList
getNonconstParameterList ()
 This is an empty stub.
 
- Public Member Functions inherited from Teuchos::ParameterListAcceptor
virtual RCP< const ParameterListgetParameterList () const
 
virtual RCP< const ParameterListgetValidParameters () const
 

Protected Member Functions

const Epetra_RowMatrixMatrix () const
 Returns a pointer to the linear system matrix.
 
int NumGlobalRows () const
 Returns the number of global rows, or -1 if Matrix() returns 0.
 
long long NumGlobalRows64 () const
 
int NumMyRows () const
 Returns the number of local rows, or -1 if Matrix() returns 0.
 
const Epetra_MapSerialMap ()
 Returns a reference to serial map (that with all elements on process 0).
 
Epetra_RowMatrixSerialMatrix ()
 Returns a reference to serial matrix (that with all rows on process 0).
 
Epetra_CrsMatrixSerialCrsMatrix ()
 
const Epetra_ImportMatrixImporter ()
 Returns a reference to the matrix importer (from row map to serial map).
 
const Epetra_ExportRhsExporter ()
 Returns a reference to the rhs exporter (from range map to serial map).
 
const Epetra_ImportSolutionImporter ()
 Returns a reference to the solution importer (to domain map from serial map).
 
int SolveSerial (Epetra_MultiVector &X, const Epetra_MultiVector &B)
 Solves the linear system, when only one process is used.
 
int SolveDistributed (Epetra_MultiVector &X, const Epetra_MultiVector &B)
 Solves the linear system, when more than one process is used.
 
int DistributedToSerial ()
 Converts a distributed matrix to serial matrix.
 
int SerialToDense ()
 Converts a serial matrix to dense format.
 
int DenseToFactored ()
 Factors the matrix using LAPACK.
 

Protected Attributes

Teuchos::RCP
< Teuchos::ParameterList
pl_
 
Teuchos::RCP< Epetra_RowMatrixSerialMatrix_
 
Teuchos::RCP< Epetra_CrsMatrixSerialCrsMatrix_
 
Teuchos::RCP< Epetra_MapSerialMap_
 
Teuchos::RCP< Epetra_ImportMatrixImporter_
 
Teuchos::RCP< Epetra_ExportRhsExporter_
 
Teuchos::RCP< Epetra_ImportSolutionImporter_
 
Epetra_SerialDenseMatrix DenseMatrix_
 Dense matrix.
 
Epetra_SerialDenseMatrix DenseLHS_
 Dense LHS.
 
Epetra_SerialDenseMatrix DenseRHS_
 Dense RHS.
 
Epetra_SerialDenseSolver DenseSolver_
 Linear problem for dense matrix and vectors.
 
bool UseTranspose_
 If true, the linear system with the transpose will be solved.
 
const Epetra_LinearProblemProblem_
 Pointer to the linear problem.
 
int MtxRedistTime_
 Quick access ids for the individual timings.
 
int MtxConvTime_
 
int VecRedistTime_
 
int SymFactTime_
 
int NumFactTime_
 
int SolveTime_
 
long long NumGlobalRows_
 
long long NumGlobalNonzeros_
 
Teuchos::RCP
< Teuchos::ParameterList
ParameterList_
 

Detailed Description

Amesos_Lapack: an interface to LAPACK.

Class Amesos_Lapack enables the solution of the distributed linear system, defined by an Epetra_LinearProblem, using LAPACK.

Amesos_Lapack stores the lineaar system matrix as an Epetra_SerialDensMatrix. The linear problem is an Epetra_SerialDenseProblem. Amesos_Lapack factorizes the matrix using DGETRF().

Date
Last updated on 16-Mar-05.
Author
Marzio Sala, 9214.

Constructor & Destructor Documentation

Amesos_Lapack::Amesos_Lapack ( const Epetra_LinearProblem LinearProblem)

Amesos_Lapack Constructor.

Creates an Amesos_Lapack instance, using an Epetra_LinearProblem, passing in an already-defined Epetra_LinearProblem object.

Note: The operator in LinearProblem must be an Epetra_RowMatrix.

References unsetParameterList().

Amesos_Lapack::~Amesos_Lapack ( void  )

Member Function Documentation

int Amesos_Lapack::GEEV ( Epetra_Vector Er,
Epetra_Vector Ei 
)

Computes the eigenvalues of the linear system matrix using DGEEV.

Parameters
Er- (Out) On processor zero only, it will contain the real component of the eigenvalues.
Ei- (Out) On processor zero only, it will contain the imaginary component of the eigenvalues.
Note
Er and Ei must have been allocated so that the local length on processor 0 equals the global size of the matrix.

References Epetra_SerialDenseMatrix::A(), DenseMatrix_, DistributedToSerial(), Epetra_LAPACK::GEEV(), Insert, Amesos_Status::IsSymbolicFactorizationOK_, Teuchos::rcp(), SerialMap(), SerialToDense(), Epetra_SerialDenseMatrix::Shape(), and SymbolicFactorization().

const Epetra_LinearProblem* Amesos_Lapack::GetProblem ( ) const
inlinevirtual

Returns the Epetra_LinearProblem.

Warning! Do not call return->SetOperator(...) to attempt to change the Epetra_Operator object (even if the new matrix has the same structure). This new operator matrix will be ignored!

Implements Amesos_BaseSolver.

References Problem_.

Referenced by Comm(), MatrixShapeOK(), and SymbolicFactorization().

bool Amesos_Lapack::MatrixShapeOK ( ) const
virtual

Returns true if the solver can handle this matrix shape.

Returns true if the matrix shape is one that the underlying sparse direct solver can handle. Classes that work only on square matrices should return false for rectangular matrices. Classes that work only on symmetric matrices whould return false for non-symmetric matrices.

Implements Amesos_BaseSolver.

References GetProblem().

int Amesos_Lapack::NumericFactorization ( )
virtual

Performs NumericFactorization on the matrix A.

In addition to performing numeric factorization on the matrix A, the call to NumericFactorization() implies that no change will be made to the underlying matrix without a subsequent call to NumericFactorization().

<br >Preconditions:

  • GetProblem().GetOperator() != 0 (return -1)
  • MatrixShapeOk(GetProblem().GetOperator()) == true (return -6)
  • The non-zero structure of the matrix should not have changed since the last call to SymbolicFactorization(). (return -2 if the number of non-zeros changes) Other changes can have arbitrary consequences.
  • The distribution of the matrix should not have changed since the last call to SymbolicFactorization()
  • The matrix should be indexed from 0 to n-1, unless the parameter "Reindex" was set to "true" prior to the call to SymbolicFactorization(). (return -3 - if caught)
  • The paremeter "Reindex" should not be set to "true" except on CrsMatrices. (return -4)
  • The paremeter "Reindex" should not be set to "true" unless Amesos was built with EpetraExt, i.e. with –enable-epetraext on the configure line. (return -4)
  • Internal errors retur -5.

<br >Postconditions:

  • Numeric Factorization will be performed (or marked to be performed) allowing Solve() to be performed correctly despite a potential change in in the matrix values (though not in the non-zero structure).
Returns
Integer error code, set to 0 if successful.

Implements Amesos_BaseSolver.

References DenseMatrix_, DenseToFactored(), DistributedToSerial(), Amesos_Status::IsNumericFactorizationOK_, Amesos_Status::IsSymbolicFactorizationOK_, Amesos_Status::NumNumericFact_, SerialToDense(), Epetra_SerialDenseMatrix::Shape(), and SymbolicFactorization().

Referenced by Solve().

void Amesos_Lapack::setParameterList ( Teuchos::RCP< Teuchos::ParameterList > const &  paramList)
virtual
int Amesos_Lapack::SetUseTranspose ( bool  UseTranspose)
inlinevirtual

If set true, X will be set to the solution of AT X = B (not A X = B)

If the implementation of this interface does not support transpose use, this method should return a value of -1.

<br >Preconditions:

<br >Postconditions:

  • The next factorization and solve will be performed with the new value of UseTranspose.
Parameters
UseTranspose– (In) If true, solve AT X = B, otherwise solve A X = B.
Returns
Integer error code, set to 0 if successful. Set to -1 if this implementation does not support transpose.

Implements Amesos_BaseSolver.

References UseTranspose_.

int Amesos_Lapack::Solve ( )
virtual

Solves A X = B (or AT x = B)

<br >Preconditions:

<br >Postconditions:

  • X will be set such that A X = B (or AT X = B), within the limits of the accuracy of the underlying solver.
Returns
Integer error code, set to 0 if successful.

Implements Amesos_BaseSolver.

References Epetra_LinearProblem::GetLHS(), Epetra_LinearProblem::GetRHS(), Amesos_Status::IsNumericFactorizationOK_, NumericFactorization(), Problem_, SolveDistributed(), and SolveSerial().

int Amesos_Lapack::SymbolicFactorization ( )
virtual

Performs SymbolicFactorization on the matrix A.

In addition to performing symbolic factorization on the matrix A, the call to SymbolicFactorization() implies that no change will be made to the non-zero structure of the underlying matrix without a subsequent call to SymbolicFactorization().

<br >Preconditions:

<br >Postconditions:

Returns
Integer error code, set to 0 if successful.

Implements Amesos_BaseSolver.

References Amesos_Time::AddTime(), Comm(), Amesos_Time::CreateTimer(), Teuchos::RCP< T >::get(), GetProblem(), Amesos_Status::IsNumericFactorizationOK_, Amesos_Status::IsSymbolicFactorizationOK_, Matrix(), Epetra_Comm::MyPID(), Epetra_Comm::NumProc(), Amesos_Status::NumSymbolicFact_, Epetra_Operator::OperatorDomainMap(), Epetra_Operator::OperatorRangeMap(), rcp(), Amesos_Time::ResetTimer(), SerialMap(), Epetra_Operator::UseTranspose(), and UseTranspose_.

Referenced by GEEV(), and NumericFactorization().


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