//////////////////////////////////////////////////////////////////////////// //////////////////////////////////////////////////////////////////////////// More...

#include <linear_solver.h>

Inheritance diagram for oomph::SuperLUSolver:

Public Types
enum	Type { Default , Serial , Distributed }
	enum type to specify the solver behaviour. Default - will employ superlu dist if more than 1 processor. Serial - will always try use superlu (serial). Distributed - will always try to use superlu dist. More...

Public Member Functions
	SuperLUSolver ()
	Constructor. Set the defaults. More...

	SuperLUSolver (const SuperLUSolver &dummy)=delete
	Broken copy constructor. More...

void	operator= (const SuperLUSolver &)=delete
	Broken assignment operator. More...

	~SuperLUSolver ()
	Destructor, clean up the stored matrices. More...

void	enable_computation_of_gradient ()
	function to enable the computation of the gradient More...

void	disable_resolve ()
	Overload disable resolve so that it cleans up memory too. More...

void	solve (Problem *const &problem_pt, DoubleVector &result)
	Solver: Takes pointer to problem and returns the results Vector which contains the solution of the linear system defined by the problem's fully assembled Jacobian and residual Vector. More...

void	solve (DoubleMatrixBase *const &matrix_pt, const DoubleVector &rhs, DoubleVector &result)
	Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data. More...

void	solve_transpose (Problem *const &problem_pt, DoubleVector &result)
	Solver: Takes pointer to problem and returns the results Vector which contains the solution of the linear system defined by the problem's fully assembled Jacobian and residual Vector. More...

void	solve_transpose (DoubleMatrixBase *const &matrix_pt, const DoubleVector &rhs, DoubleVector &result)
	Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data. More...

void	resolve (const DoubleVector &rhs, DoubleVector &result)
	Resolve the system defined by the last assembled jacobian and the specified rhs vector if resolve has been enabled. Note: returns the global result Vector. More...

void	resolve_transpose (const DoubleVector &rhs, DoubleVector &result)
	Resolve the (transposed) system defined by the last assembled Jacobian and the specified rhs vector if resolve has been enabled. More...

void	enable_doc_stats ()
	Enable documentation of solver statistics. More...

void	disable_doc_stats ()
	Disable documentation of solver statistics. More...

double	jacobian_setup_time () const
	returns the time taken to assemble the jacobian matrix and residual vector More...

virtual double	linear_solver_solution_time () const
	return the time taken to solve the linear system (needs to be overloaded for each linear solver) More...

void	factorise (DoubleMatrixBase *const &matrix_pt)
	Do the factorisation stage Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data. More...

void	backsub (const DoubleVector &rhs, DoubleVector &result)
	Do the backsubstitution for SuperLU solver Note: returns the global result Vector. More...

void	backsub_transpose (const DoubleVector &rhs, DoubleVector &result)
	Do the back-substitution for transposed system of the SuperLU solver Note: Returns the global result Vector. More...

void	clean_up_memory ()
	Clean up the memory allocated by the solver. More...

void	set_solver_type (const Type &t)
	Specify the solve type. Either default, serial or distributed. See enum SuperLU_solver_type for more details. More...

void	use_compressed_row_for_superlu_serial ()
	Use the compressed row format in superlu serial. More...

void	use_compressed_column_for_superlu_serial ()
	Use the compressed column format in superlu serial. More...

void	enable_delete_matrix_data_in_superlu_dist ()
	Set Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be deleted the flag can be set to true to reduce the amount of memory required, and the matrix data will be wiped using its clean_up_memory() function. Default value is false. More...

void	disable_delete_matrix_data_in_superlu_dist ()
	Unset Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be. More...

void	enable_row_and_col_permutations_in_superlu_dist ()
	Set flag so that SuperLU_DIST is allowed to permute matrix rows and columns during factorisation. This is the default for SuperLU_DIST, and can lead to significantly faster solves. More...

void	disable_row_and_col_permutations_in_superlu_dist ()
	Set flag so that SuperLU_DIST is not allowed to permute matrix rows and columns during factorisation. More...

void	use_global_solve_in_superlu_dist ()
	Calling this method will ensure that when the problem based solve interface is used, a global (serial) jacobian will be assembled. Note: calling this function will delete any distributed solve data. More...

void	use_distributed_solve_in_superlu_dist ()
	Calling this method will ensure that when the problem based solve interface is used, a distributed jacobian will be assembled. Note: calling this function will delete any global solve data. More...

double	get_memory_usage_for_lu_factors ()
	How much memory do the LU factors take up? In bytes. More...

double	get_total_needed_memory ()
	How much memory was allocated by SuperLU? In bytes. More...

Public Member Functions inherited from oomph::LinearSolver
	LinearSolver ()
	Empty constructor, initialise the member data. More...

	LinearSolver (const LinearSolver &dummy)=delete
	Broken copy constructor. More...

void	operator= (const LinearSolver &)=delete
	Broken assignment operator. More...

virtual	~LinearSolver ()
	Empty virtual destructor. More...

void	enable_doc_time ()
	Enable documentation of solve times. More...

void	disable_doc_time ()
	Disable documentation of solve times. More...

bool	is_doc_time_enabled () const
	Is documentation of solve times enabled? More...

bool	is_resolve_enabled () const
	Boolean flag indicating if resolves are enabled. More...

virtual void	enable_resolve ()
	Enable resolve (i.e. store matrix and/or LU decomposition, say) Virtual so it can be overloaded to perform additional tasks. More...

virtual void	solve (DoubleMatrixBase *const &matrix_pt, const Vector< double > &rhs, Vector< double > &result)
	Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. More...

virtual void	solve_transpose (DoubleMatrixBase *const &matrix_pt, const Vector< double > &rhs, Vector< double > &result)
	Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. More...

void	disable_computation_of_gradient ()
	function to disable the computation of the gradient required for the globally convergent Newton method More...

void	reset_gradient ()
	function to reset the size of the gradient before each Newton solve More...

void	get_gradient (DoubleVector &gradient)
	function to access the gradient, provided it has been computed More...

Public Member Functions inherited from oomph::DistributableLinearAlgebraObject
	DistributableLinearAlgebraObject ()
	Default constructor - create a distribution. More...

	DistributableLinearAlgebraObject (const DistributableLinearAlgebraObject &matrix)=delete
	Broken copy constructor. More...

void	operator= (const DistributableLinearAlgebraObject &)=delete
	Broken assignment operator. More...

virtual	~DistributableLinearAlgebraObject ()
	Destructor. More...

LinearAlgebraDistribution *	distribution_pt () const
	access to the LinearAlgebraDistribution More...

unsigned	nrow () const
	access function to the number of global rows. More...

unsigned	nrow_local () const
	access function for the num of local rows on this processor. More...

unsigned	nrow_local (const unsigned &p) const
	access function for the num of local rows on this processor. More...

unsigned	first_row () const
	access function for the first row on this processor More...

unsigned	first_row (const unsigned &p) const
	access function for the first row on this processor More...

bool	distributed () const
	distribution is serial or distributed More...

bool	distribution_built () const
	if the communicator_pt is null then the distribution is not setup then false is returned, otherwise return true More...

void	build_distribution (const LinearAlgebraDistribution *const dist_pt)
	setup the distribution of this distributable linear algebra object More...

void	build_distribution (const LinearAlgebraDistribution &dist)
	setup the distribution of this distributable linear algebra object More...

Static Public Attributes
static bool	Suppress_incorrect_rhs_distribution_warning_in_resolve
	Static flag that determines whether the warning about incorrect distribution of RHSs will be printed or not. More...

Private Member Functions
void	factorise_serial (DoubleMatrixBase *const &matrix_pt)
	factorise method for SuperLU (serial) More...

void	backsub_serial (const DoubleVector &rhs, DoubleVector &result)
	backsub method for SuperLU (serial) More...

void	backsub_transpose_serial (const DoubleVector &rhs, DoubleVector &result)
	backsub method for SuperLU (serial) More...

void	factorise_distributed (DoubleMatrixBase *const &matrix_pt)
	factorise method for SuperLU Dist More...

void	backsub_distributed (const DoubleVector &rhs, DoubleVector &result)
	backsub method for SuperLU Dist More...

void	backsub_transpose_distributed (const DoubleVector &rhs, DoubleVector &result)
	backsub method for SuperLU Dist More...

Private Attributes
double	Jacobian_setup_time
	Jacobian setup time. More...

double	Solution_time
	Solution time. More...

bool	Suppress_solve
	Suppress solve? More...

bool	Doc_stats
	Set to true to output statistics (false by default). More...

Type	Solver_type
	the solver type. see SuperLU_solver_type for details. More...

bool	Using_dist
	boolean flag indicating whether superlu dist is being used More...

void *	Serial_f_factors
	Storage for the LU factors as required by SuperLU. More...

int	Serial_info
	Info flag for the SuperLU solver. More...

unsigned long	Serial_n_dof
	The number of unknowns in the linear system. More...

int	Serial_sign_of_determinant_of_matrix
	Sign of the determinant of the matrix. More...

bool	Serial_compressed_row_flag
	Use compressed row version? More...

bool	Dist_use_global_solver
	Flag that determines whether the MPIProblem based solve function uses the global or distributed version of SuperLU_DIST (default value is false). More...

bool	Dist_global_solve_data_allocated
	Flag is true if solve data has been generated for a global matrix. More...

bool	Dist_distributed_solve_data_allocated
	Flag is true if solve data has been generated for distributed matrix. More...

void *	Dist_solver_data_pt
	Storage for the LU factors and other data required by SuperLU. More...

int	Dist_nprow
	Number of rows for the process grid. More...

int	Dist_npcol
	Number of columns for the process grid. More...

int	Dist_info
	Info flag for the SuperLU solver. More...

bool	Dist_allow_row_and_col_permutations
	If true then SuperLU_DIST is allowed to permute matrix rows and columns during factorisation. This is the default for SuperLU_DIST, and can lead to significantly faster solves, but has been known to fail, hence the default value is 0. More...

bool	Dist_delete_matrix_data
	Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be deleted the flag can be set to true to reduce the amount of memory required, and the matrix data will be wiped using its clean_up_memory() function. Default value is false. More...

double *	Dist_value_pt
	Pointer for storage of the matrix values required by SuperLU_DIST. More...

int *	Dist_index_pt
	Pointer for storage of matrix rows or column indices required by SuperLU_DIST. More...

int *	Dist_start_pt
	Pointers for storage of matrix column or row starts. More...

Additional Inherited Members
Protected Member Functions inherited from oomph::DistributableLinearAlgebraObject
void	clear_distribution ()
	clear the distribution of this distributable linear algebra object More...

Protected Attributes inherited from oomph::LinearSolver
bool	Enable_resolve
	Boolean that indicates whether the matrix (or its factors, in the case of direct solver) should be stored so that the resolve function can be used. More...

bool	Doc_time
	Boolean flag that indicates whether the time taken. More...

bool	Compute_gradient
	flag that indicates whether the gradient required for the globally convergent Newton method should be computed or not More...

bool	Gradient_has_been_computed
	flag that indicates whether the gradient was computed or not More...

DoubleVector	Gradient_for_glob_conv_newton_solve
	DoubleVector storing the gradient for the globally convergent Newton method. More...

Detailed Description

//////////////////////////////////////////////////////////////////////////// ////////////////////////////////////////////////////////////////////////////

SuperLU Project Solver class. This is a combined wrapper for both SuperLU and SuperLU Dist. See http://crd.lbl.gov/~xiaoye/SuperLU/ Default Behaviour: If this solver is distributed over more than one processor then SuperLU Dist is used. Member data naming convention: member data associated with the SuperLU Dist solver begins Dist_... and member data associated with the serial SuperLU solver begins Serial_... .

Definition at line 485 of file linear_solver.h.

Member Enumeration Documentation

◆ Type

enum oomph::SuperLUSolver::Type

enum type to specify the solver behaviour. Default - will employ superlu dist if more than 1 processor. Serial - will always try use superlu (serial). Distributed - will always try to use superlu dist.

Enumerator
Default
Serial
Distributed

Definition at line 492 of file linear_solver.h.

Constructor & Destructor Documentation

◆ SuperLUSolver() [1/2]

oomph::SuperLUSolver::SuperLUSolver ( )

inline

Constructor. Set the defaults.

Definition at line 500 of file linear_solver.h.

References Default, Dist_allow_row_and_col_permutations, Dist_delete_matrix_data, Dist_distributed_solve_data_allocated, Dist_global_solve_data_allocated, Dist_index_pt, Dist_solver_data_pt, Dist_start_pt, Dist_use_global_solver, Dist_value_pt, Doc_stats, Serial_compressed_row_flag, Serial_n_dof, Serial_sign_of_determinant_of_matrix, Solver_type, Suppress_solve, and Using_dist.

◆ SuperLUSolver() [2/2]

oomph::SuperLUSolver::SuperLUSolver ( const SuperLUSolver & dummy )

delete

Broken copy constructor.

◆ ~SuperLUSolver()

oomph::SuperLUSolver::~SuperLUSolver ( )

inline

Destructor, clean up the stored matrices.

Definition at line 534 of file linear_solver.h.

References clean_up_memory().

Member Function Documentation

◆ backsub()

void oomph::SuperLUSolver::backsub	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

Do the backsubstitution for SuperLU solver Note: returns the global result Vector.

Do the backsubstitution for SuperLUSolver. Note - this method performs no paranoid checks - these are all performed in solve(...) and resolve(...)

Definition at line 2299 of file linear_solver.cc.

References backsub_distributed(), backsub_serial(), and Using_dist.

Referenced by resolve(), and solve().

◆ backsub_distributed()

void oomph::SuperLUSolver::backsub_distributed	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

private

backsub method for SuperLU Dist

Do the backsubstitution for SuperLU solver. Note - this method performs no paranoid checks - these are all performed in solve(...) and resolve(...)

Definition at line 2347 of file linear_solver.cc.

References oomph::LinearAlgebraDistribution::built(), Dist_distributed_solve_data_allocated, Dist_global_solve_data_allocated, Dist_info, Dist_npcol, Dist_nprow, Dist_solver_data_pt, oomph::DistributableLinearAlgebraObject::distribution_built(), oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LeakCheckNames::doc(), Doc_stats, oomph::LinearAlgebraDistribution::nrow(), oomph::superlu_dist_distributed_matrix(), oomph::superlu_dist_global_matrix(), Suppress_incorrect_rhs_distribution_warning_in_resolve, and oomph::DoubleVector::values_pt().

Referenced by backsub().

◆ backsub_serial()

void oomph::SuperLUSolver::backsub_serial	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

private

backsub method for SuperLU (serial)

Do the backsubstitution for SuperLU.

Definition at line 2510 of file linear_solver.cc.

References oomph::DoubleVector::built(), oomph::LinearAlgebraDistribution::distributed(), oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LeakCheckNames::doc(), Doc_stats, i, oomph::DistributableLinearAlgebraObject::nrow(), Serial_compressed_row_flag, Serial_f_factors, Serial_info, Serial_n_dof, oomph::superlu(), and oomph::DoubleVector::values_pt().

Referenced by backsub().

◆ backsub_transpose()

void oomph::SuperLUSolver::backsub_transpose	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

Do the back-substitution for transposed system of the SuperLU solver Note: Returns the global result Vector.

Do the backsubstitution of the transposed system for SuperLUSolver. Note - this method performs no paranoid checks - these are all performed in solve(...) and resolve(...)

Definition at line 2319 of file linear_solver.cc.

References backsub_transpose_distributed(), backsub_transpose_serial(), and Using_dist.

Referenced by resolve_transpose(), and solve_transpose().

◆ backsub_transpose_distributed()

void oomph::SuperLUSolver::backsub_transpose_distributed	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

private

backsub method for SuperLU Dist

Do the backsubstitution for SuperLU solver. Note - this method performs no paranoid checks - these are all performed in solve(...) and resolve(...)

Definition at line 2490 of file linear_solver.cc.

Referenced by backsub_transpose().

◆ backsub_transpose_serial()

void oomph::SuperLUSolver::backsub_transpose_serial	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

private

backsub method for SuperLU (serial)

Do the backsubstitution for SuperLU.

Definition at line 2600 of file linear_solver.cc.

References oomph::DoubleVector::built(), oomph::LinearAlgebraDistribution::distributed(), oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LeakCheckNames::doc(), Doc_stats, i, oomph::DistributableLinearAlgebraObject::nrow(), Serial_compressed_row_flag, Serial_f_factors, Serial_info, Serial_n_dof, oomph::superlu(), and oomph::DoubleVector::values_pt().

Referenced by backsub_transpose().

◆ clean_up_memory()

void oomph::SuperLUSolver::clean_up_memory ( )

virtual

Clean up the memory allocated by the solver.

Clean up the memory.

Reimplemented from oomph::LinearSolver.

Definition at line 2690 of file linear_solver.cc.

References oomph::DistributableLinearAlgebraObject::clear_distribution(), Dist_distributed_solve_data_allocated, Dist_global_solve_data_allocated, Dist_index_pt, Dist_info, Dist_npcol, Dist_nprow, Dist_solver_data_pt, Dist_start_pt, Dist_value_pt, oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LeakCheckNames::doc(), Doc_stats, i, oomph::LinearAlgebraDistribution::nrow(), Serial_compressed_row_flag, Serial_f_factors, Serial_info, Serial_n_dof, oomph::superlu(), oomph::superlu_dist_distributed_matrix(), and oomph::superlu_dist_global_matrix().

Referenced by oomph::SuperLUPreconditioner::clean_up_memory(), disable_resolve(), factorise(), factorise_distributed(), factorise_serial(), set_solver_type(), solve(), solve_transpose(), use_distributed_solve_in_superlu_dist(), use_global_solve_in_superlu_dist(), and ~SuperLUSolver().

◆ disable_delete_matrix_data_in_superlu_dist()

void oomph::SuperLUSolver::disable_delete_matrix_data_in_superlu_dist ( )

inline

Unset Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be.

Definition at line 686 of file linear_solver.h.

References Dist_delete_matrix_data.

◆ disable_doc_stats()

void oomph::SuperLUSolver::disable_doc_stats ( )

inline

Disable documentation of solver statistics.

Definition at line 609 of file linear_solver.h.

References Doc_stats.

Referenced by oomph::SuperLUPreconditioner::disable_doc_stats(), and oomph::SuperLUPreconditioner::SuperLUPreconditioner().

◆ disable_resolve()

void oomph::SuperLUSolver::disable_resolve ( )

inlinevirtual

Overload disable resolve so that it cleans up memory too.

Reimplemented from oomph::LinearSolver.

Definition at line 549 of file linear_solver.h.

References clean_up_memory(), and oomph::LinearSolver::disable_resolve().

◆ disable_row_and_col_permutations_in_superlu_dist()

void oomph::SuperLUSolver::disable_row_and_col_permutations_in_superlu_dist ( )

inline

Set flag so that SuperLU_DIST is not allowed to permute matrix rows and columns during factorisation.

Definition at line 701 of file linear_solver.h.

References Dist_allow_row_and_col_permutations.

◆ enable_computation_of_gradient()

void oomph::SuperLUSolver::enable_computation_of_gradient ( )

inlinevirtual

function to enable the computation of the gradient

Reimplemented from oomph::LinearSolver.

Definition at line 540 of file linear_solver.h.

References oomph::LinearSolver::Compute_gradient.

◆ enable_delete_matrix_data_in_superlu_dist()

void oomph::SuperLUSolver::enable_delete_matrix_data_in_superlu_dist ( )

inline

Set Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be deleted the flag can be set to true to reduce the amount of memory required, and the matrix data will be wiped using its clean_up_memory() function. Default value is false.

Definition at line 678 of file linear_solver.h.

References Dist_delete_matrix_data.

◆ enable_doc_stats()

void oomph::SuperLUSolver::enable_doc_stats ( )

inline

Enable documentation of solver statistics.

Definition at line 603 of file linear_solver.h.

References Doc_stats.

Referenced by oomph::SuperLUPreconditioner::enable_doc_stats().

◆ enable_row_and_col_permutations_in_superlu_dist()

void oomph::SuperLUSolver::enable_row_and_col_permutations_in_superlu_dist ( )

inline

Set flag so that SuperLU_DIST is allowed to permute matrix rows and columns during factorisation. This is the default for SuperLU_DIST, and can lead to significantly faster solves.

Definition at line 694 of file linear_solver.h.

References Dist_allow_row_and_col_permutations.

◆ factorise()

void oomph::SuperLUSolver::factorise ( DoubleMatrixBase *const & matrix_pt )

Do the factorisation stage Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data.

LU decompose the matrix addressed by matrix_pt by using the SuperLU solver. The resulting matrix factors are stored internally.

Definition at line 1827 of file linear_solver.cc.

References clean_up_memory(), oomph::LinearAlgebraDistribution::communicator_pt(), Default, Distributed, oomph::DistributableLinearAlgebraObject::distribution_pt(), factorise_distributed(), factorise_serial(), oomph::MPI_Helpers::mpi_has_been_initialised(), Solver_type, and Using_dist.

Referenced by oomph::SuperLUPreconditioner::setup(), solve(), and solve_transpose().

◆ factorise_distributed()

void oomph::SuperLUSolver::factorise_distributed ( DoubleMatrixBase *const & matrix_pt )

private

factorise method for SuperLU Dist

LU decompose the matrix addressed by matrix_pt using the SuperLU_DIST solver. The resulting matrix factors are stored internally.

Definition at line 1872 of file linear_solver.cc.

References oomph::DistributableLinearAlgebraObject::build_distribution(), oomph::CRDoubleMatrix::built(), clean_up_memory(), oomph::CCMatrix< T >::clean_up_memory(), oomph::CRDoubleMatrix::clear(), oomph::CRDoubleMatrix::column_index(), oomph::CCMatrix< T >::column_start(), oomph::MPI_Helpers::communicator_pt(), oomph::LinearAlgebraDistribution::communicator_pt(), Dist_allow_row_and_col_permutations, Dist_delete_matrix_data, Dist_distributed_solve_data_allocated, Dist_global_solve_data_allocated, Dist_index_pt, Dist_info, Dist_npcol, Dist_nprow, Dist_solver_data_pt, Dist_start_pt, Dist_value_pt, oomph::DistributableLinearAlgebraObject::distributed(), oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LeakCheckNames::doc(), Doc_stats, oomph::DistributableLinearAlgebraObject::first_row(), i, oomph::DoubleMatrixBase::ncol(), oomph::SparseMatrix< T, MATRIX_TYPE >::nnz(), oomph::CRDoubleMatrix::nnz(), oomph::LinearAlgebraDistribution::nrow(), oomph::DistributableLinearAlgebraObject::nrow(), oomph::CRDoubleMatrix::nrow(), oomph::CCDoubleMatrix::nrow(), oomph::DoubleMatrixBase::nrow(), oomph::DistributableLinearAlgebraObject::nrow_local(), oomph::CCMatrix< T >::row_index(), oomph::CRDoubleMatrix::row_start(), oomph::superlu_cr_to_cc(), oomph::superlu_dist_distributed_matrix(), oomph::superlu_dist_global_matrix(), oomph::SparseMatrix< T, MATRIX_TYPE >::value(), and oomph::CRDoubleMatrix::value().

Referenced by factorise().

◆ factorise_serial()

void oomph::SuperLUSolver::factorise_serial ( DoubleMatrixBase *const & matrix_pt )

private

factorise method for SuperLU (serial)

LU decompose the matrix addressed by matrix_pt by using the SuperLU solver. The resulting matrix factors are stored internally.

Definition at line 2168 of file linear_solver.cc.

References clean_up_memory(), oomph::CRDoubleMatrix::column_index(), oomph::CCMatrix< T >::column_start(), oomph::DistributableLinearAlgebraObject::distributed(), oomph::LeakCheckNames::doc(), Doc_stats, i, oomph::DoubleMatrixBase::ncol(), oomph::SparseMatrix< T, MATRIX_TYPE >::nnz(), oomph::CRDoubleMatrix::nnz(), oomph::DoubleMatrixBase::nrow(), oomph::CCMatrix< T >::row_index(), oomph::CRDoubleMatrix::row_start(), Serial_compressed_row_flag, Serial_f_factors, Serial_info, Serial_n_dof, Serial_sign_of_determinant_of_matrix, oomph::CumulativeTimings::start(), oomph::superlu(), oomph::SparseMatrix< T, MATRIX_TYPE >::value(), and oomph::CRDoubleMatrix::value().

Referenced by factorise().

◆ get_memory_usage_for_lu_factors()

double oomph::SuperLUSolver::get_memory_usage_for_lu_factors ( )

How much memory do the LU factors take up? In bytes.

How much memory do the LU factors take up? In bytes NOTE: This has been scraped from dQuerySpace(...) in dmemory.c in external_src/oomph_superlu_4.3.

Definition at line 758 of file linear_solver.cc.

References Dist_solver_data_pt, Distributed, oomph::get_lu_factor_memory_usage_in_bytes(), oomph::get_lu_factor_memory_usage_in_bytes_dist(), Serial_f_factors, and Solver_type.

Referenced by oomph::SuperLUPreconditioner::get_memory_usage_for_lu_factors(), and oomph::SuperLUPreconditioner::get_memory_usage_for_superlu().

◆ get_total_needed_memory()

double oomph::SuperLUSolver::get_total_needed_memory ( )

How much memory was allocated by SuperLU? In bytes.

How much memory was used in total? In bytes NOTE: This has been scraped from dQuerySpace(...) in dmemory.c in external_src/oomph_superlu_4.3.

Definition at line 786 of file linear_solver.cc.

References Dist_solver_data_pt, Distributed, oomph::get_total_memory_usage_in_bytes(), oomph::get_total_memory_usage_in_bytes_dist(), Serial_f_factors, and Solver_type.

Referenced by oomph::SuperLUPreconditioner::get_memory_usage_for_superlu(), oomph::SuperLUPreconditioner::get_total_memory_needed_for_superlu(), solve(), and solve_transpose().

◆ jacobian_setup_time()

double oomph::SuperLUSolver::jacobian_setup_time ( ) const

inlinevirtual

returns the time taken to assemble the jacobian matrix and residual vector

Reimplemented from oomph::LinearSolver.

Definition at line 616 of file linear_solver.h.

References Jacobian_setup_time.

◆ linear_solver_solution_time()

virtual double oomph::SuperLUSolver::linear_solver_solution_time ( ) const

inlinevirtual

return the time taken to solve the linear system (needs to be overloaded for each linear solver)

Reimplemented from oomph::LinearSolver.

Definition at line 623 of file linear_solver.h.

References Solution_time.

◆ operator=()

void oomph::SuperLUSolver::operator= ( const SuperLUSolver & )

delete

Broken assignment operator.

◆ resolve()

void oomph::SuperLUSolver::resolve	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

virtual

Resolve the system defined by the last assembled jacobian and the specified rhs vector if resolve has been enabled. Note: returns the global result Vector.

Resolve the system for a given RHS.

Reimplemented from oomph::LinearSolver.

Definition at line 1776 of file linear_solver.cc.

References backsub(), oomph::TimingHelpers::convert_secs_to_formatted_string(), oomph::LinearSolver::Doc_time, oomph::DistributableLinearAlgebraObject::nrow(), oomph::oomph_info, Solution_time, and oomph::TimingHelpers::timer().

Referenced by oomph::SuperLUPreconditioner::preconditioner_solve().

◆ resolve_transpose()

void oomph::SuperLUSolver::resolve_transpose	(	const DoubleVector &	rhs,
		DoubleVector &	result
	)

virtual

Resolve the (transposed) system defined by the last assembled Jacobian and the specified rhs vector if resolve has been enabled.

Resolve the (transposed) system for a given RHS.

Reimplemented from oomph::LinearSolver.

Definition at line 1800 of file linear_solver.cc.

References backsub_transpose(), oomph::TimingHelpers::convert_secs_to_formatted_string(), oomph::LinearSolver::Doc_time, oomph::DistributableLinearAlgebraObject::nrow(), oomph::oomph_info, Solution_time, and oomph::TimingHelpers::timer().

Referenced by oomph::SuperLUPreconditioner::preconditioner_solve_transpose().

◆ set_solver_type()

void oomph::SuperLUSolver::set_solver_type ( const Type & t )

inline

Specify the solve type. Either default, serial or distributed. See enum SuperLU_solver_type for more details.

Definition at line 646 of file linear_solver.h.

References clean_up_memory(), Solver_type, and t.

◆ solve() [1/2]

void oomph::SuperLUSolver::solve	(	DoubleMatrixBase *const &	matrix_pt,
		const DoubleVector &	rhs,
		DoubleVector &	result
	)

virtual

Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data.

Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. Problem pointer defaults to NULL and can be omitted. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data.

Reimplemented from oomph::LinearSolver.

Definition at line 1074 of file linear_solver.cc.

References backsub(), oomph::DistributableLinearAlgebraObject::build_distribution(), oomph::DoubleVector::built(), clean_up_memory(), oomph::TimingHelpers::convert_secs_to_formatted_string(), oomph::LinearAlgebraDistribution::distributed(), oomph::DistributableLinearAlgebraObject::distribution_pt(), Doc_stats, oomph::LinearSolver::Doc_time, oomph::LinearSolver::Enable_resolve, factorise(), get_total_needed_memory(), oomph::DoubleMatrixBase::ncol(), oomph::CRDoubleMatrix::nnz(), oomph::DistributableLinearAlgebraObject::nrow(), oomph::CRDoubleMatrix::nrow(), oomph::DoubleMatrixBase::nrow(), oomph::oomph_info, Solution_time, and oomph::TimingHelpers::timer().

◆ solve() [2/2]

void oomph::SuperLUSolver::solve	(	Problem *const &	problem_pt,
		DoubleVector &	result
	)

virtual

Solver: Takes pointer to problem and returns the results Vector which contains the solution of the linear system defined by the problem's fully assembled Jacobian and residual Vector.

Implements oomph::LinearSolver.

Definition at line 814 of file linear_solver.cc.

References oomph::DoubleVector::build(), oomph::DistributableLinearAlgebraObject::build_distribution(), oomph::DoubleVector::built(), clean_up_memory(), oomph::Problem::communicator_pt(), oomph::LinearSolver::Compute_gradient, oomph::TimingHelpers::convert_secs_to_formatted_string(), Default, Dist_delete_matrix_data, Dist_use_global_solver, Distributed, oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LinearSolver::Doc_time, oomph::Problem::get_jacobian(), oomph::LinearSolver::Gradient_for_glob_conv_newton_solve, oomph::LinearSolver::Gradient_has_been_computed, Jacobian_setup_time, oomph::CRDoubleMatrix::multiply_transpose(), oomph::CCDoubleMatrix::multiply_transpose(), oomph::Problem::ndof(), oomph::oomph_info, oomph::DoubleVector::redistribute(), Serial_compressed_row_flag, Serial_sign_of_determinant_of_matrix, oomph::Problem::sign_of_jacobian(), Solver_type, Suppress_solve, and oomph::TimingHelpers::timer().

◆ solve_transpose() [1/2]

void oomph::SuperLUSolver::solve_transpose	(	DoubleMatrixBase *const &	matrix_pt,
		const DoubleVector &	rhs,
		DoubleVector &	result
	)

virtual

Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data.

Linear-algebra-type solver: Takes pointer to a matrix and rhs vector and returns the solution of the linear system. Problem pointer defaults to NULL and can be omitted. The function returns the global result Vector. Note: if Delete_matrix_data is true the function matrix_pt->clean_up_memory() will be used to wipe the matrix data.

Reimplemented from oomph::LinearSolver.

Definition at line 1558 of file linear_solver.cc.

References backsub_transpose(), oomph::DistributableLinearAlgebraObject::build_distribution(), oomph::DoubleVector::built(), clean_up_memory(), oomph::TimingHelpers::convert_secs_to_formatted_string(), oomph::LinearAlgebraDistribution::distributed(), oomph::DistributableLinearAlgebraObject::distribution_pt(), Doc_stats, oomph::LinearSolver::Doc_time, oomph::LinearSolver::Enable_resolve, factorise(), get_total_needed_memory(), oomph::DoubleMatrixBase::ncol(), oomph::CRDoubleMatrix::nnz(), oomph::DistributableLinearAlgebraObject::nrow(), oomph::CRDoubleMatrix::nrow(), oomph::DoubleMatrixBase::nrow(), oomph::oomph_info, Solution_time, and oomph::TimingHelpers::timer().

◆ solve_transpose() [2/2]

void oomph::SuperLUSolver::solve_transpose	(	Problem *const &	problem_pt,
		DoubleVector &	result
	)

virtual

Solver: Takes pointer to problem and returns the results Vector which contains the solution of the linear system defined by the problem's fully assembled Jacobian and residual Vector.

Reimplemented from oomph::LinearSolver.

Definition at line 1297 of file linear_solver.cc.

References oomph::DoubleVector::build(), oomph::DistributableLinearAlgebraObject::build_distribution(), oomph::DoubleVector::built(), clean_up_memory(), oomph::Problem::communicator_pt(), oomph::LinearSolver::Compute_gradient, oomph::TimingHelpers::convert_secs_to_formatted_string(), Default, Dist_delete_matrix_data, Dist_use_global_solver, Distributed, oomph::DistributableLinearAlgebraObject::distribution_pt(), oomph::LinearSolver::Doc_time, oomph::Problem::get_jacobian(), oomph::LinearSolver::Gradient_for_glob_conv_newton_solve, oomph::LinearSolver::Gradient_has_been_computed, Jacobian_setup_time, oomph::CRDoubleMatrix::multiply_transpose(), oomph::CCDoubleMatrix::multiply_transpose(), oomph::Problem::ndof(), oomph::oomph_info, oomph::DoubleVector::redistribute(), Serial_compressed_row_flag, Serial_sign_of_determinant_of_matrix, oomph::Problem::sign_of_jacobian(), Solver_type, Suppress_solve, and oomph::TimingHelpers::timer().

◆ use_compressed_column_for_superlu_serial()

void oomph::SuperLUSolver::use_compressed_column_for_superlu_serial ( )

inline

Use the compressed column format in superlu serial.

Definition at line 662 of file linear_solver.h.

References Serial_compressed_row_flag.

◆ use_compressed_row_for_superlu_serial()

void oomph::SuperLUSolver::use_compressed_row_for_superlu_serial ( )

inline

Use the compressed row format in superlu serial.

Definition at line 656 of file linear_solver.h.

References Serial_compressed_row_flag.

◆ use_distributed_solve_in_superlu_dist()

void oomph::SuperLUSolver::use_distributed_solve_in_superlu_dist ( )

inline

Calling this method will ensure that when the problem based solve interface is used, a distributed jacobian will be assembled. Note: calling this function will delete any global solve data.

Definition at line 723 of file linear_solver.h.

References clean_up_memory(), and Dist_use_global_solver.

◆ use_global_solve_in_superlu_dist()

void oomph::SuperLUSolver::use_global_solve_in_superlu_dist ( )

inline

Calling this method will ensure that when the problem based solve interface is used, a global (serial) jacobian will be assembled. Note: calling this function will delete any distributed solve data.

Definition at line 710 of file linear_solver.h.

References clean_up_memory(), and Dist_use_global_solver.

Member Data Documentation

◆ Dist_allow_row_and_col_permutations

bool oomph::SuperLUSolver::Dist_allow_row_and_col_permutations

private

If true then SuperLU_DIST is allowed to permute matrix rows and columns during factorisation. This is the default for SuperLU_DIST, and can lead to significantly faster solves, but has been known to fail, hence the default value is 0.

Definition at line 841 of file linear_solver.h.

Referenced by disable_row_and_col_permutations_in_superlu_dist(), enable_row_and_col_permutations_in_superlu_dist(), factorise_distributed(), and SuperLUSolver().

◆ Dist_delete_matrix_data

bool oomph::SuperLUSolver::Dist_delete_matrix_data

private

Delete_matrix_data flag. SuperLU_dist needs its own copy of the input matrix, therefore a copy must be made if any matrix used with this solver is to be preserved. If the input matrix can be deleted the flag can be set to true to reduce the amount of memory required, and the matrix data will be wiped using its clean_up_memory() function. Default value is false.

Definition at line 849 of file linear_solver.h.

Referenced by disable_delete_matrix_data_in_superlu_dist(), enable_delete_matrix_data_in_superlu_dist(), factorise_distributed(), solve(), solve_transpose(), and SuperLUSolver().

◆ Dist_distributed_solve_data_allocated

bool oomph::SuperLUSolver::Dist_distributed_solve_data_allocated

private

Flag is true if solve data has been generated for distributed matrix.

Definition at line 823 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), factorise_distributed(), and SuperLUSolver().

◆ Dist_global_solve_data_allocated

bool oomph::SuperLUSolver::Dist_global_solve_data_allocated

private

Flag is true if solve data has been generated for a global matrix.

Definition at line 820 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), factorise_distributed(), and SuperLUSolver().

◆ Dist_index_pt

int* oomph::SuperLUSolver::Dist_index_pt

private

Pointer for storage of matrix rows or column indices required by SuperLU_DIST.

Definition at line 856 of file linear_solver.h.

Referenced by clean_up_memory(), factorise_distributed(), and SuperLUSolver().

◆ Dist_info

int oomph::SuperLUSolver::Dist_info

private

Info flag for the SuperLU solver.

Definition at line 835 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), and factorise_distributed().

◆ Dist_npcol

int oomph::SuperLUSolver::Dist_npcol

private

Number of columns for the process grid.

Definition at line 832 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), and factorise_distributed().

◆ Dist_nprow

int oomph::SuperLUSolver::Dist_nprow

private

Number of rows for the process grid.

Definition at line 829 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), and factorise_distributed().

◆ Dist_solver_data_pt

void* oomph::SuperLUSolver::Dist_solver_data_pt

private

Storage for the LU factors and other data required by SuperLU.

Definition at line 826 of file linear_solver.h.

Referenced by backsub_distributed(), clean_up_memory(), factorise_distributed(), get_memory_usage_for_lu_factors(), get_total_needed_memory(), and SuperLUSolver().

◆ Dist_start_pt

int* oomph::SuperLUSolver::Dist_start_pt

private

Pointers for storage of matrix column or row starts.

Definition at line 860 of file linear_solver.h.

Referenced by clean_up_memory(), factorise_distributed(), and SuperLUSolver().

◆ Dist_use_global_solver

bool oomph::SuperLUSolver::Dist_use_global_solver

private

Flag that determines whether the MPIProblem based solve function uses the global or distributed version of SuperLU_DIST (default value is false).

Definition at line 817 of file linear_solver.h.

Referenced by solve(), solve_transpose(), SuperLUSolver(), use_distributed_solve_in_superlu_dist(), and use_global_solve_in_superlu_dist().

◆ Dist_value_pt

double* oomph::SuperLUSolver::Dist_value_pt

private

Pointer for storage of the matrix values required by SuperLU_DIST.

Definition at line 852 of file linear_solver.h.

Referenced by clean_up_memory(), factorise_distributed(), and SuperLUSolver().

◆ Doc_stats

bool oomph::SuperLUSolver::Doc_stats

private

Set to true to output statistics (false by default).

Definition at line 770 of file linear_solver.h.

Referenced by backsub_distributed(), backsub_serial(), backsub_transpose_serial(), clean_up_memory(), disable_doc_stats(), enable_doc_stats(), factorise_distributed(), factorise_serial(), solve(), solve_transpose(), and SuperLUSolver().

◆ Jacobian_setup_time

double oomph::SuperLUSolver::Jacobian_setup_time

private

Jacobian setup time.

Definition at line 761 of file linear_solver.h.

Referenced by jacobian_setup_time(), solve(), and solve_transpose().

◆ Serial_compressed_row_flag

bool oomph::SuperLUSolver::Serial_compressed_row_flag

private

Use compressed row version?

Definition at line 794 of file linear_solver.h.

Referenced by backsub_serial(), backsub_transpose_serial(), clean_up_memory(), factorise_serial(), solve(), solve_transpose(), SuperLUSolver(), use_compressed_column_for_superlu_serial(), and use_compressed_row_for_superlu_serial().

◆ Serial_f_factors

void* oomph::SuperLUSolver::Serial_f_factors

private

Storage for the LU factors as required by SuperLU.

Definition at line 782 of file linear_solver.h.

Referenced by backsub_serial(), backsub_transpose_serial(), clean_up_memory(), factorise_serial(), get_memory_usage_for_lu_factors(), and get_total_needed_memory().

◆ Serial_info

int oomph::SuperLUSolver::Serial_info

private

Info flag for the SuperLU solver.

Definition at line 785 of file linear_solver.h.

Referenced by backsub_serial(), backsub_transpose_serial(), clean_up_memory(), and factorise_serial().

◆ Serial_n_dof

unsigned long oomph::SuperLUSolver::Serial_n_dof

private

The number of unknowns in the linear system.

Definition at line 788 of file linear_solver.h.

Referenced by backsub_serial(), backsub_transpose_serial(), clean_up_memory(), factorise_serial(), and SuperLUSolver().

◆ Serial_sign_of_determinant_of_matrix

int oomph::SuperLUSolver::Serial_sign_of_determinant_of_matrix

private

Sign of the determinant of the matrix.

Definition at line 791 of file linear_solver.h.

Referenced by factorise_serial(), solve(), solve_transpose(), and SuperLUSolver().

◆ Solution_time

double oomph::SuperLUSolver::Solution_time

private

Solution time.

Definition at line 764 of file linear_solver.h.

Referenced by linear_solver_solution_time(), resolve(), resolve_transpose(), solve(), and solve_transpose().

◆ Solver_type

Type oomph::SuperLUSolver::Solver_type

private

the solver type. see SuperLU_solver_type for details.

Definition at line 773 of file linear_solver.h.

Referenced by factorise(), get_memory_usage_for_lu_factors(), get_total_needed_memory(), set_solver_type(), solve(), solve_transpose(), and SuperLUSolver().

◆ Suppress_incorrect_rhs_distribution_warning_in_resolve

bool oomph::SuperLUSolver::Suppress_incorrect_rhs_distribution_warning_in_resolve

static

Initial value:

=

false

Static flag that determines whether the warning about incorrect distribution of RHSs will be printed or not.

Static warning to suppress warnings about incorrect distribution of RHS vector. Default is false.

Definition at line 811 of file linear_solver.h.

Referenced by backsub_distributed().

◆ Suppress_solve

bool oomph::SuperLUSolver::Suppress_solve

private

Suppress solve?

Definition at line 767 of file linear_solver.h.

Referenced by solve(), solve_transpose(), and SuperLUSolver().

◆ Using_dist

bool oomph::SuperLUSolver::Using_dist

private

boolean flag indicating whether superlu dist is being used

Definition at line 776 of file linear_solver.h.

Referenced by backsub(), backsub_transpose(), factorise(), and SuperLUSolver().

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

Public Types

Public Member Functions

Static Public Attributes

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Member Enumeration Documentation

◆ Type

Constructor & Destructor Documentation

◆ SuperLUSolver() [1/2]

◆ SuperLUSolver() [2/2]

◆ ~SuperLUSolver()

Member Function Documentation

◆ backsub()

◆ backsub_distributed()

◆ backsub_serial()

◆ backsub_transpose()

◆ backsub_transpose_distributed()

◆ backsub_transpose_serial()

◆ clean_up_memory()

◆ disable_delete_matrix_data_in_superlu_dist()

◆ disable_doc_stats()

◆ disable_resolve()

◆ disable_row_and_col_permutations_in_superlu_dist()

◆ enable_computation_of_gradient()

◆ enable_delete_matrix_data_in_superlu_dist()

◆ enable_doc_stats()

◆ enable_row_and_col_permutations_in_superlu_dist()

◆ factorise()

◆ factorise_distributed()

◆ factorise_serial()

◆ get_memory_usage_for_lu_factors()

◆ get_total_needed_memory()

◆ jacobian_setup_time()

◆ linear_solver_solution_time()

◆ operator=()

◆ resolve()

◆ resolve_transpose()

◆ set_solver_type()

◆ solve() [1/2]

◆ solve() [2/2]

◆ solve_transpose() [1/2]

◆ solve_transpose() [2/2]

◆ use_compressed_column_for_superlu_serial()

◆ use_compressed_row_for_superlu_serial()

◆ use_distributed_solve_in_superlu_dist()

◆ use_global_solve_in_superlu_dist()

Member Data Documentation

◆ Dist_allow_row_and_col_permutations

◆ Dist_delete_matrix_data

◆ Dist_distributed_solve_data_allocated

◆ Dist_global_solve_data_allocated

◆ Dist_index_pt

◆ Dist_info

◆ Dist_npcol

◆ Dist_nprow

◆ Dist_solver_data_pt

◆ Dist_start_pt

◆ Dist_use_global_solver

◆ Dist_value_pt

◆ Doc_stats

◆ Jacobian_setup_time

◆ Serial_compressed_row_flag

◆ Serial_f_factors

◆ Serial_info

◆ Serial_n_dof

◆ Serial_sign_of_determinant_of_matrix

◆ Solution_time

◆ Solver_type

◆ Suppress_incorrect_rhs_distribution_warning_in_resolve

◆ Suppress_solve

◆ Using_dist