The GEMSLR information on each level, contains the solver for B and the low-rank information for S. More...

#include <parallel_gemslr.hpp>

Public Member Functions
int	Clear ()
	Free the current level structure, set everything to 0. More...

	ParallelGemslrLevelClass ()
	The constructor of ParallelGemslrLevelClass, set everything to 0. More...

	~ParallelGemslrLevelClass ()
	The destructor of ParallelGemslrLevelClass. More...

	ParallelGemslrLevelClass (const ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &str)
	The copy constructor of ParallelGemslrLevelClass. More...

	ParallelGemslrLevelClass (ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &&str)
	The move constructor of ParallelGemslrLevelClass. More...

ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &	operator= (const ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &str)
	The operator= of ParallelGemslrLevelClass. More...

ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &	operator= (ParallelGemslrLevelClass< MatrixType, VectorType, DataType > &&str)
	The operator= of ParallelGemslrLevelClass. More...

int	GetNumNonzeros (long int &nnz_bsolver, long int &nnz_lr)
	Get the number of nonzeros in the low-rank correction and the ILU factorization on this level. More...

Public Attributes
parallel_log	_parlog
	The parallel log data structure. More...

int	_lrc
	The size of low-rank correction on this level. More...

int	_ncomps
	Number of subdomains on this level. More...

std::vector< CsrMatrixClass< DataType > >	_B_mat_v
	The B matrix on this level. C matrix if this is the last level. More...

MatrixType	_E_mat
	The E matrix on this level. More...

MatrixType	_F_mat
	The F matrix on this level. More...

MatrixType	_A_mat
	The A matrix on this level. More...

MatrixType	_C_mat
	The C matrix on this level. More...

MatrixType	_S_mat
	The S matrix on this level. More...

int	_nI
	Number of interior nodes on the top level. More...

ParallelGemslrEBFCMatrixClass< MatrixType, VectorType, DataType >	_EBFC
	The EBFC matrix on this level. More...

SolverClass< CsrMatrixClass< DataType >, SequentialVectorClass< DataType >, DataType > **	_B_precond
	The preconditioners for B matrix. More...

SolverClass< CsrMatrixClass< DataType >, SequentialVectorClass< DataType >, DataType > **	_B_solver
	The solvers for B matrix. More...

DenseMatrixClass< DataType >	_Hk
	The H matrix for the low-rank correction WHW' on this level. More...

DenseMatrixClass< DataType >	_Wk
	The W matrix for the low-rank correction WHW' on this level. More...

DenseMatrixClass< DataType >	_WHk
	The WH matrix for the low-rank correction (WH)W' on this level. More...

DenseMatrixClass< DataType >	_cHk
	The H matrix for the low-rank correction WHW' on the last level. More...

DenseMatrixClass< DataType >	_cWk
	The W matrix for the low-rank correction WHW' on the last level. More...

DenseMatrixClass< DataType >	_cWHk
	The WH matrix for the low-rank correction (WH)W' on the last level. More...

IntVectorClass< int >	_pperm
	The row permutation vector. More...

IntVectorClass< int >	_qperm
	The column permutation vector. More...

CommunicationHelperClass	_comm_helper
	The communication helper. More...

SequentialVectorClass< DataType >	_work_vector
	Temp vector on this level. More...

int	_work_vector_unit_length
	The unit length of the work vector. More...

VectorType	_x_temp
	Temp vector, length equal to the size of this level, used to setup the Ptr for this level. More...

VectorType	_xlr_temp
	Temp vector. More...

VectorType	_xlr1_temp
	Temp vector. More...

VectorType	_xlr2_temp
	Temp vector. More...

VectorType	_xlr1_temp_h
	Temp vector. More...

VectorType	_xlr2_temp_h
	Temp vector. More...

VectorType	_sol_temp
	The temp vector for permuted x. More...

VectorType	_rhs_temp
	The temp vector for permuted rhs. More...

VectorType	_sol2_temp
	The temp vector for permuted x. More...

VectorType	_rhs2_temp
	The temp vector for permuted rhs. More...

VectorType	_sol3_temp
	The temp vector for permuted x for the upper level. More...

VectorType	_rhs3_temp
	The temp vector for permuted rhs for the upper level. More...

Detailed Description

template<class MatrixType, class VectorType, typename DataType>
class pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >

The GEMSLR information on each level, contains the solver for B and the low-rank information for S. VectorType is the type of the vector. DataType is the data type.

Constructor & Destructor Documentation

◆ ParallelGemslrLevelClass() [1/3]

template<class MatrixType , class VectorType , typename DataType >

template pargemslr::precond_gemslrlevel_csr_par_complexd::ParallelGemslrLevelClass ( )

The constructor of ParallelGemslrLevelClass, set everything to 0.

◆ ~ParallelGemslrLevelClass()

template<class MatrixType , class VectorType , typename DataType >

template pargemslr::precond_gemslrlevel_csr_par_complexd::~ParallelGemslrLevelClass ( )

The destructor of ParallelGemslrLevelClass, simply call the free function.

◆ ParallelGemslrLevelClass() [2/3]

template<class MatrixType , class VectorType , typename DataType >

pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::ParallelGemslrLevelClass ( const ParallelGemslrLevelClass< MatrixType, VectorType, DataType > & str )

The copy constructor of ParallelGemslrLevelClass.

◆ ParallelGemslrLevelClass() [3/3]

template<class MatrixType , class VectorType , typename DataType >

pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::ParallelGemslrLevelClass ( ParallelGemslrLevelClass< MatrixType, VectorType, DataType > && str )

The move constructor of ParallelGemslrLevelClass.

Member Function Documentation

◆ Clear()

template<class MatrixType , class VectorType , typename DataType >

template int pargemslr::precond_gemslrlevel_csr_par_complexd::Clear ( )

Free the current level structure, set everything to 0.

Returns: return GEMSLR_SUCESS or error information.

◆ GetNumNonzeros()

template<class MatrixType , class VectorType , typename DataType >

template int pargemslr::precond_gemslrlevel_csr_par_complexd::GetNumNonzeros	(	long int &	nnz_bsolver,
		long int &	nnz_lr
	)

Get the number of nonzeros in the low-rank correction and the ILU factorization on this level.
Note that this is the sequential version, int would be enough. No need to use the long int.

Parameters

[out]	nnz_ilu	The nnz for the ILU factorization.
[out]	nnz_lr	The nnz for the low-rank correction.

Returns: return nnz_ilu+nnz_lr.

◆ operator=() [1/2]

template<class MatrixType , class VectorType , typename DataType >

ParallelGemslrLevelClass< MatrixType, VectorType, DataType > & pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::operator= ( const ParallelGemslrLevelClass< MatrixType, VectorType, DataType > & str )

The operator= of ParallelGemslrLevelClass.

◆ operator=() [2/2]

template<class MatrixType , class VectorType , typename DataType >

ParallelGemslrLevelClass< MatrixType, VectorType, DataType > & pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::operator= ( ParallelGemslrLevelClass< MatrixType, VectorType, DataType > && str )

The operator= of ParallelGemslrLevelClass.

Member Data Documentation

◆ _A_mat

template<class MatrixType , class VectorType , typename DataType >

MatrixType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_A_mat

The A matrix on this level.

◆ _B_mat_v

template<class MatrixType , class VectorType , typename DataType >

std::vector<CsrMatrixClass<DataType> > pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_B_mat_v

The B matrix on this level. C matrix if this is the last level.

◆ _B_precond

template<class MatrixType , class VectorType , typename DataType >

SolverClass<CsrMatrixClass<DataType>, SequentialVectorClass<DataType>, DataType>** pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_B_precond

The preconditioners for B matrix.

◆ _B_solver

template<class MatrixType , class VectorType , typename DataType >

SolverClass<CsrMatrixClass<DataType>, SequentialVectorClass<DataType>, DataType>** pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_B_solver

The solvers for B matrix.

◆ _C_mat

template<class MatrixType , class VectorType , typename DataType >

MatrixType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_C_mat

The C matrix on this level.

◆ _cHk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_cHk

The H matrix for the low-rank correction W*H*W' on the last level.

◆ _comm_helper

template<class MatrixType , class VectorType , typename DataType >

CommunicationHelperClass pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_comm_helper

The communication helper.

◆ _cWHk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_cWHk

The WH matrix for the low-rank correction (W*H)*W' on the last level.

◆ _cWk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_cWk

The W matrix for the low-rank correction W*H*W' on the last level.

◆ _E_mat

template<class MatrixType , class VectorType , typename DataType >

MatrixType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_E_mat

The E matrix on this level.

◆ _EBFC

template<class MatrixType , class VectorType , typename DataType >

ParallelGemslrEBFCMatrixClass<MatrixType, VectorType, DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_EBFC

The EBFC matrix on this level.

◆ _F_mat

template<class MatrixType , class VectorType , typename DataType >

MatrixType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_F_mat

The F matrix on this level.

◆ _Hk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_Hk

The H matrix for the low-rank correction W*H*W' on this level.

◆ _lrc

template<class MatrixType , class VectorType , typename DataType >

int pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_lrc

The size of low-rank correction on this level.

◆ _ncomps

template<class MatrixType , class VectorType , typename DataType >

int pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_ncomps

Number of subdomains on this level.

◆ _nI

template<class MatrixType , class VectorType , typename DataType >

int pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_nI

Number of interior nodes on the top level.

◆ _parlog

template<class MatrixType , class VectorType , typename DataType >

parallel_log pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_parlog

The parallel log data structure.

◆ _pperm

template<class MatrixType , class VectorType , typename DataType >

IntVectorClass<int> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_pperm

The row permutation vector.

◆ _qperm

template<class MatrixType , class VectorType , typename DataType >

IntVectorClass<int> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_qperm

The column permutation vector.

Note: For non-symmetric reordering.

◆ _rhs2_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_rhs2_temp

The temp vector for permuted rhs.

◆ _rhs3_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_rhs3_temp

The temp vector for permuted rhs for the upper level.

◆ _rhs_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_rhs_temp

The temp vector for permuted rhs.

◆ _S_mat

template<class MatrixType , class VectorType , typename DataType >

MatrixType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_S_mat

The S matrix on this level.

◆ _sol2_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_sol2_temp

The temp vector for permuted x.

◆ _sol3_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_sol3_temp

The temp vector for permuted x for the upper level.

◆ _sol_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_sol_temp

The temp vector for permuted x.

◆ _WHk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_WHk

The WH matrix for the low-rank correction (W*H)*W' on this level.

◆ _Wk

template<class MatrixType , class VectorType , typename DataType >

DenseMatrixClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_Wk

The W matrix for the low-rank correction W*H*W' on this level.

◆ _work_vector

template<class MatrixType , class VectorType , typename DataType >

SequentialVectorClass<DataType> pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_work_vector

Temp vector on this level.

◆ _work_vector_unit_length

template<class MatrixType , class VectorType , typename DataType >

int pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_work_vector_unit_length

The unit length of the work vector.

◆ _x_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_x_temp

Temp vector, length equal to the size of this level, used to setup the Ptr for this level.

◆ _xlr1_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_xlr1_temp

Temp vector.

◆ _xlr1_temp_h

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_xlr1_temp_h

Temp vector.

◆ _xlr2_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_xlr2_temp

Temp vector.

◆ _xlr2_temp_h

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_xlr2_temp_h

Temp vector.

◆ _xlr_temp

template<class MatrixType , class VectorType , typename DataType >

VectorType pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >::_xlr_temp

Temp vector.

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

SRC/preconditioners/parallel_gemslr.hpp
SRC/preconditioners/parallel_gemslr.cpp

Public Member Functions

Public Attributes

Detailed Description

template<class MatrixType, class VectorType, typename DataType> class pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >

Constructor & Destructor Documentation

◆ ParallelGemslrLevelClass() [1/3]

◆ ~ParallelGemslrLevelClass()

◆ ParallelGemslrLevelClass() [2/3]

◆ ParallelGemslrLevelClass() [3/3]

Member Function Documentation

◆ Clear()

◆ GetNumNonzeros()

◆ operator=() [1/2]

◆ operator=() [2/2]

Member Data Documentation

◆ _A_mat

◆ _B_mat_v

◆ _B_precond

◆ _B_solver

◆ _C_mat

◆ _cHk

◆ _comm_helper

◆ _cWHk

◆ _cWk

◆ _E_mat

◆ _EBFC

◆ _F_mat

◆ _Hk

◆ _lrc

◆ _ncomps

◆ _nI

◆ _parlog

◆ _pperm

◆ _qperm

◆ _rhs2_temp

◆ _rhs3_temp

◆ _rhs_temp

◆ _S_mat

◆ _sol2_temp

◆ _sol3_temp

◆ _sol_temp

◆ _WHk

◆ _Wk

◆ _work_vector

◆ _work_vector_unit_length

◆ _x_temp

◆ _xlr1_temp

◆ _xlr1_temp_h

◆ _xlr2_temp

◆ _xlr2_temp_h

◆ _xlr_temp

template<class MatrixType, class VectorType, typename DataType>
class pargemslr::ParallelGemslrLevelClass< MatrixType, VectorType, DataType >