MFEM v4.7.0
Finite element discretization library
Loading...
Searching...
No Matches
mfem::Ginkgo::GinkgoPreconditioner Class Reference

#include <ginkgo.hpp>

Inheritance diagram for mfem::Ginkgo::GinkgoPreconditioner:
[legend]
Collaboration diagram for mfem::Ginkgo::GinkgoPreconditioner:
[legend]

Public Member Functions

 GinkgoPreconditioner (GinkgoExecutor &exec)
 
virtual ~GinkgoPreconditioner ()=default
 
virtual void SetOperator (const Operator &op)
 
virtual void Mult (const Vector &x, Vector &y) const
 
const std::shared_ptr< gko::LinOpFactory > GetFactory () const
 
const std::shared_ptr< gko::LinOp > GetGeneratedPreconditioner () const
 
bool HasGeneratedPreconditioner () const
 
- Public Member Functions inherited from mfem::Solver
 Solver (int s=0, bool iter_mode=false)
 Initialize a square Solver with size s.
 
 Solver (int h, int w, bool iter_mode=false)
 Initialize a Solver with height h and width w.
 
- Public Member Functions inherited from mfem::Operator
void InitTVectors (const Operator *Po, const Operator *Ri, const Operator *Pi, Vector &x, Vector &b, Vector &X, Vector &B) const
 Initializes memory for true vectors of linear system.
 
 Operator (int s=0)
 Construct a square Operator with given size s (default 0).
 
 Operator (int h, int w)
 Construct an Operator with the given height (output size) and width (input size).
 
int Height () const
 Get the height (size of output) of the Operator. Synonym with NumRows().
 
int NumRows () const
 Get the number of rows (size of output) of the Operator. Synonym with Height().
 
int Width () const
 Get the width (size of input) of the Operator. Synonym with NumCols().
 
int NumCols () const
 Get the number of columns (size of input) of the Operator. Synonym with Width().
 
virtual MemoryClass GetMemoryClass () const
 Return the MemoryClass preferred by the Operator.
 
virtual void MultTranspose (const Vector &x, Vector &y) const
 Action of the transpose operator: y=A^t(x). The default behavior in class Operator is to generate an error.
 
virtual void AddMult (const Vector &x, Vector &y, const real_t a=1.0) const
 Operator application: y+=A(x) (default) or y+=a*A(x).
 
virtual void AddMultTranspose (const Vector &x, Vector &y, const real_t a=1.0) const
 Operator transpose application: y+=A^t(x) (default) or y+=a*A^t(x).
 
virtual void ArrayMult (const Array< const Vector * > &X, Array< Vector * > &Y) const
 Operator application on a matrix: Y=A(X).
 
virtual void ArrayMultTranspose (const Array< const Vector * > &X, Array< Vector * > &Y) const
 Action of the transpose operator on a matrix: Y=A^t(X).
 
virtual void ArrayAddMult (const Array< const Vector * > &X, Array< Vector * > &Y, const real_t a=1.0) const
 Operator application on a matrix: Y+=A(X) (default) or Y+=a*A(X).
 
virtual void ArrayAddMultTranspose (const Array< const Vector * > &X, Array< Vector * > &Y, const real_t a=1.0) const
 Operator transpose application on a matrix: Y+=A^t(X) (default) or Y+=a*A^t(X).
 
virtual OperatorGetGradient (const Vector &x) const
 Evaluate the gradient operator at the point x. The default behavior in class Operator is to generate an error.
 
virtual void AssembleDiagonal (Vector &diag) const
 Computes the diagonal entries into diag. Typically, this operation only makes sense for linear Operators. In some cases, only an approximation of the diagonal is computed.
 
virtual const OperatorGetProlongation () const
 Prolongation operator from linear algebra (linear system) vectors, to input vectors for the operator. NULL means identity.
 
virtual const OperatorGetRestriction () const
 Restriction operator from input vectors for the operator to linear algebra (linear system) vectors. NULL means identity.
 
virtual const OperatorGetOutputProlongation () const
 Prolongation operator from linear algebra (linear system) vectors, to output vectors for the operator. NULL means identity.
 
virtual const OperatorGetOutputRestrictionTranspose () const
 Transpose of GetOutputRestriction, directly available in this form to facilitate matrix-free RAP-type operators.
 
virtual const OperatorGetOutputRestriction () const
 Restriction operator from output vectors for the operator to linear algebra (linear system) vectors. NULL means identity.
 
void FormLinearSystem (const Array< int > &ess_tdof_list, Vector &x, Vector &b, Operator *&A, Vector &X, Vector &B, int copy_interior=0)
 Form a constrained linear system using a matrix-free approach.
 
void FormRectangularLinearSystem (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, Vector &x, Vector &b, Operator *&A, Vector &X, Vector &B)
 Form a column-constrained linear system using a matrix-free approach.
 
virtual void RecoverFEMSolution (const Vector &X, const Vector &b, Vector &x)
 Reconstruct a solution vector x (e.g. a GridFunction) from the solution X of a constrained linear system obtained from Operator::FormLinearSystem() or Operator::FormRectangularLinearSystem().
 
void FormSystemOperator (const Array< int > &ess_tdof_list, Operator *&A)
 Return in A a parallel (on truedofs) version of this square operator.
 
void FormRectangularSystemOperator (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, Operator *&A)
 Return in A a parallel (on truedofs) version of this rectangular operator (including constraints).
 
void FormDiscreteOperator (Operator *&A)
 Return in A a parallel (on truedofs) version of this rectangular operator.
 
void PrintMatlab (std::ostream &out, int n, int m=0) const
 Prints operator with input size n and output size m in Matlab format.
 
virtual void PrintMatlab (std::ostream &out) const
 Prints operator in Matlab format.
 
virtual ~Operator ()
 Virtual destructor.
 
Type GetType () const
 Return the type ID of the Operator class.
 

Protected Attributes

std::shared_ptr< gko::LinOpFactory > precond_gen
 
std::shared_ptr< gko::LinOp > generated_precond
 
std::shared_ptr< gko::Executor > executor
 
bool has_generated_precond
 
- Protected Attributes inherited from mfem::Operator
int height
 Dimension of the output / number of rows in the matrix.
 
int width
 Dimension of the input / number of columns in the matrix.
 

Additional Inherited Members

- Public Types inherited from mfem::Operator
enum  DiagonalPolicy { DIAG_ZERO , DIAG_ONE , DIAG_KEEP }
 Defines operator diagonal policy upon elimination of rows and/or columns. More...
 
enum  Type {
  ANY_TYPE , MFEM_SPARSEMAT , Hypre_ParCSR , PETSC_MATAIJ ,
  PETSC_MATIS , PETSC_MATSHELL , PETSC_MATNEST , PETSC_MATHYPRE ,
  PETSC_MATGENERIC , Complex_Operator , MFEM_ComplexSparseMat , Complex_Hypre_ParCSR ,
  Complex_DenseMat , MFEM_Block_Matrix , MFEM_Block_Operator
}
 Enumeration defining IDs for some classes derived from Operator. More...
 
- Public Attributes inherited from mfem::Solver
bool iterative_mode
 If true, use the second argument of Mult() as an initial guess.
 
- Protected Member Functions inherited from mfem::Operator
void FormConstrainedSystemOperator (const Array< int > &ess_tdof_list, ConstrainedOperator *&Aout)
 see FormSystemOperator()
 
void FormRectangularConstrainedSystemOperator (const Array< int > &trial_tdof_list, const Array< int > &test_tdof_list, RectangularConstrainedOperator *&Aout)
 see FormRectangularSystemOperator()
 
OperatorSetupRAP (const Operator *Pi, const Operator *Po)
 Returns RAP Operator of this, using input/output Prolongation matrices Pi corresponds to "P", Po corresponds to "Rt".
 

Detailed Description

This class forms the base class for all of Ginkgo's preconditioners. The various derived classes only take the additional data that is specific to them. The entire collection of preconditioners that Ginkgo implements is available at the Ginkgo documentation and manual pages, https://ginkgo-project.github.io/ginkgo/doc/develop.

Definition at line 576 of file ginkgo.hpp.

Constructor & Destructor Documentation

◆ GinkgoPreconditioner()

mfem::Ginkgo::GinkgoPreconditioner::GinkgoPreconditioner ( GinkgoExecutor & exec)

Constructor.

The exec defines the paradigm where the solution is computed. Ginkgo currently supports four different executor types:

  • OmpExecutor specifies that the data should be stored and the associated operations executed on an OpenMP-supporting device (e.g. host CPU);
  • CudaExecutor specifies that the data should be stored and the operations executed on the NVIDIA GPU accelerator;
  • HipExecutor specifies that the data should be stored and the operations executed on the GPU accelerator using HIP;
  • ReferenceExecutor executes a non-optimized reference implementation, which can be used to debug the library.

Definition at line 1080 of file ginkgo.cpp.

◆ ~GinkgoPreconditioner()

virtual mfem::Ginkgo::GinkgoPreconditioner::~GinkgoPreconditioner ( )
virtualdefault

Destructor.

Member Function Documentation

◆ GetFactory()

const std::shared_ptr< gko::LinOpFactory > mfem::Ginkgo::GinkgoPreconditioner::GetFactory ( ) const
inline

Return a pointer to the LinOpFactory that will generate the preconditioner with the parameters set through the specific constructor.

Definition at line 621 of file ginkgo.hpp.

◆ GetGeneratedPreconditioner()

const std::shared_ptr< gko::LinOp > mfem::Ginkgo::GinkgoPreconditioner::GetGeneratedPreconditioner ( ) const
inline

Return a pointer to the generated preconditioner for a specific matrix (that has previously been set with SetOperator).

Definition at line 630 of file ginkgo.hpp.

◆ HasGeneratedPreconditioner()

bool mfem::Ginkgo::GinkgoPreconditioner::HasGeneratedPreconditioner ( ) const
inline

Return whether this GinkgoPreconditioner object has an explicitly- generated preconditioner, built for a specific matrix.

Definition at line 639 of file ginkgo.hpp.

◆ Mult()

void mfem::Ginkgo::GinkgoPreconditioner::Mult ( const Vector & x,
Vector & y ) const
virtual

Apply the preconditioner to input vector x, with out y.

Implements mfem::Operator.

Definition at line 1089 of file ginkgo.cpp.

◆ SetOperator()

void mfem::Ginkgo::GinkgoPreconditioner::SetOperator ( const Operator & op)
virtual

Generate the preconditioner for the given matrix op, which must be of MFEM SparseMatrix type. Calling this function is only required when creating a preconditioner for use with another MFEM solver; to use with a Ginkgo solver, get the LinOpFactory pointer through GetFactory() and pass to the Ginkgo solver constructor.

Implements mfem::Solver.

Reimplemented in mfem::Ginkgo::MFEMPreconditioner.

Definition at line 1123 of file ginkgo.cpp.

Member Data Documentation

◆ executor

std::shared_ptr<gko::Executor> mfem::Ginkgo::GinkgoPreconditioner::executor
protected

The execution paradigm in Ginkgo. The choices are between gko::OmpExecutor, gko::CudaExecutor and gko::ReferenceExecutor and more details can be found in Ginkgo's documentation.

Definition at line 662 of file ginkgo.hpp.

◆ generated_precond

std::shared_ptr<gko::LinOp> mfem::Ginkgo::GinkgoPreconditioner::generated_precond
protected

Generated Ginkgo preconditioner for a specific matrix, created through SetOperator(), or a wrapped MFEM preconditioner. Must exist to use Mult().

Definition at line 655 of file ginkgo.hpp.

◆ has_generated_precond

bool mfem::Ginkgo::GinkgoPreconditioner::has_generated_precond
protected

Whether or not we have generated a specific preconditioner for a matrix.

Definition at line 668 of file ginkgo.hpp.

◆ precond_gen

std::shared_ptr<gko::LinOpFactory> mfem::Ginkgo::GinkgoPreconditioner::precond_gen
protected

The Ginkgo generated solver factory object.

Definition at line 648 of file ginkgo.hpp.


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