#include <MatrixVectorHomogen.h>

Public Member Functions
	MatrixVectorHomogen (AdaptiveSparseGrid &grid, MultiLevelAdaptiveSparseGrid &mgrid)
	Constructor for MatrixVectorHomogen.

	MatrixVectorHomogen (AdaptiveSparseGrid &grid, MultiLevelAdaptiveSparseGrid &mgrid, bool symmetry)
	Constructor for MatrixVectorHomogen.

	MatrixVectorHomogen (AdaptiveSparseGrid &grid, MultiLevelAdaptiveSparseGrid &mgrid, bool symmetry, bool mpi_on)
	Constructor for MatrixVectorHomogen.

template<class StencilClass >
void	multiplication (VectorSparseG &prew, VectorSparseG &Ax, StencilClass &stencilClass)

template<class Problem >
void	casefunction_mpi (VectorSparseG &prew, VectorSparseG &Ax, Problem &stencilClass)
	Distribute the cases with dynamic mpi_scheduling.

Public Attributes
bool	option_symmetry = false

bool	option_mpi = false

int	stencil_count =0

Private Member Functions
void	calcNodal (VectorSparseG &prew, Depth &depth)

void	calcNodal (VectorSparseG &nodal_u, VectorSparseG &prew, Depth &depth)

template<class StencilClass >
void	caseFunction (VectorSparseG &prew, bool *restrictions, StencilClass &stencil)

template<class Problem >
void	caseFunction_LocalStiffnessMatrix (VectorSparseG &prew, bool *restrictions, Problem &localStiffnessMatrixAllDepths)

template<class StencilClass >
void	applyStencilGlobal (const VectorSparseG &input, VectorSparseG &output, StencilClass &stencil, Depth &T)

template<class StencilClass >
void	applyStencilGlobal_symmetry (VectorSparseG &input, VectorSparseG &output, StencilClass &stencil, Depth &T)

template<class StencilClass >
void	applyStencilGlobal_MPI (VectorSparseG &input, VectorSparseG &output, StencilClass &stencil, Depth &T)

template<class StencilClass >
void	applyStencilGlobal_MPI_2 (VectorSparseG &input, VectorSparseG &output, StencilClass &stencil, Depth &T)

template<class StencilClass >
void	applyStencilGlobal_MPI_3 (VectorSparseG &input, VectorSparseG &output, StencilClass &stencil, Depth &T)

template<class StencilClass >
double	applyStencilLocal (const IndexDimension &Index, const VectorSparseG &input, StencilClass &stencil, Depth &T)

void	prolongation1D_inplace (VectorSparseG &values, int &t, int &d)

void	restriction1D_inverted_inplace (const VectorSparseG &values, int t, int d)

void	ConvertToPrewavelet2 (VectorSparseG &Ax_nodal, VectorSparseG &Ax, Depth &T)

void	master (int num_workers, int total_tasks)

void	master_start (int num_workers)

void	master_distribute ()

void	master_end_distribute ()

template<class StencilClass >
void	worker (VectorSparseG &prew, VectorSparseG &Ax, StencilClass &stencilClass)

void	intToBinary (int num, bool binary[])

Detailed Description

This class provides a Matrix-Vector-Multiplication on Dirichlet-SparseGrids

Constructor & Destructor Documentation

◆ MatrixVectorHomogen() [1/3]

MatrixVectorHomogen::MatrixVectorHomogen	(	AdaptiveSparseGrid &	grid,
		MultiLevelAdaptiveSparseGrid &	mgrid
	)

inline

Constructor for MatrixVectorHomogen.

Parameters

grid	AdaptiveSparseGrid, make sure that the grid is completed, i.e. fullfills Definition of locally adaptive sparse grid
mgrid	locally adaptive full subgrids

◆ MatrixVectorHomogen() [2/3]

MatrixVectorHomogen::MatrixVectorHomogen	(	AdaptiveSparseGrid &	grid,
		MultiLevelAdaptiveSparseGrid &	mgrid,
		bool	symmetry
	)

inline

Constructor for MatrixVectorHomogen.

Parameters

grid	AdaptiveSparseGrid, make sure that the grid is completed, i.e. fullfills Definition of locally adaptive sparse grid
mgrid	locally adaptive full subgrids
symmetry	make use of symmetry of the bilinear form

◆ MatrixVectorHomogen() [3/3]

MatrixVectorHomogen::MatrixVectorHomogen	(	AdaptiveSparseGrid &	grid,
		MultiLevelAdaptiveSparseGrid &	mgrid,
		bool	symmetry,
		bool	mpi_on
	)

inline

Constructor for MatrixVectorHomogen.

Parameters

grid	AdaptiveSparseGrid, make sure that the grid is completed, i.e. fullfills Definition of locally adaptive sparse grid
mgrid	locally adaptive full subgrids
symmetry	make use of symmetry of the bilinear form
mpi_on	distribute local stiffness matrices with MPI

Member Function Documentation

◆ applyStencilGlobal()

template<class StencilClass >

void MatrixVectorHomogen::applyStencilGlobal	(	const VectorSparseG &	input,
		VectorSparseG &	output,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on all grid points with depth less or equal to Depth T

Template Parameters

StencilClass

Parameters

input	Nodal values
output	a(input, v_p)
stencil
T

Here is the caller graph for this function:

◆ applyStencilGlobal_MPI()

template<class StencilClass >

void MatrixVectorHomogen::applyStencilGlobal_MPI	(	VectorSparseG &	input,
		VectorSparseG &	output,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on all grid points with depth less or equal to Depth T. Make use of Symmetry of the bilinearform by calculating local stiffness matrices on cells. Distribute the local stiffness matrices with MPI. (No SpeedUp yet.)

Template Parameters

StencilClass

Parameters

input	Nodal values
output	a(input, v_p)
stencil
T

◆ applyStencilGlobal_MPI_2()

template<class StencilClass >

void MatrixVectorHomogen::applyStencilGlobal_MPI_2	(	VectorSparseG &	input,
		VectorSparseG &	output,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on all grid points with depth less or equal to Depth T. Make use of Symmetry of the bilinearform by calculating local stiffness matrices on cells. Distribute the local stiffness matrices with MPI. (No SpeedUp yet.)

Template Parameters

StencilClass

Parameters

input	Nodal values
output	a(input, v_p)
stencil
T

◆ applyStencilGlobal_MPI_3()

template<class StencilClass >

void MatrixVectorHomogen::applyStencilGlobal_MPI_3	(	VectorSparseG &	input,
		VectorSparseG &	output,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on all grid points with depth less or equal to Depth T. Make use of Symmetry of the bilinearform by calculating local stiffness matrices on cells. Distribute the local stiffness matrices with MPI. (No SpeedUp yet.)

Template Parameters

StencilClass

Parameters

input	Nodal values
output	a(input, v_p)
stencil
T

Here is the caller graph for this function:

◆ applyStencilGlobal_symmetry()

template<class StencilClass >

void MatrixVectorHomogen::applyStencilGlobal_symmetry	(	VectorSparseG &	input,
		VectorSparseG &	output,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on all grid points with depth less or equal to Depth T. Make use of Symmetry of the bilinearform by calculating local stiffness matrices on cells.

Template Parameters

StencilClass

Parameters

input	Nodal values
output	a(input, v_p)
stencil
T

Here is the caller graph for this function:

◆ applyStencilLocal()

template<class StencilClass >

double MatrixVectorHomogen::applyStencilLocal	(	const IndexDimension &	Index,
		const VectorSparseG &	input,
		StencilClass &	stencil,
		Depth &	T
	)

private

Apply a stencil on a given grid point (Index) w.r.t. Depth T. I.e. a(u,v_p) with u = nodal, p = Index.

Template Parameters

StencilClass

Parameters

Index	grid point
input	nodal values
stencil
T

Returns

◆ calcNodal() [1/2]

void MatrixVectorHomogen::calcNodal	(	VectorSparseG &	nodal_u,
		VectorSparseG &	prew,
		Depth &	depth
	)

private

Calculates nodal values of a function given of prewavelets (prew) on a fixed level (depth)

Parameters

prew
depth

◆ calcNodal() [2/2]

void MatrixVectorHomogen::calcNodal	(	VectorSparseG &	prew,
		Depth &	depth
	)

private

Calculates nodal values of a function given of prewavelets (prew) on a fixed level (depth)

Parameters

prew
depth

◆ caseFunction()

template<class StencilClass >

void MatrixVectorHomogen::caseFunction	(	VectorSparseG &	prew,
		bool *	restrictions,
		StencilClass &	stencil
	)

private

Calculates one of the 2^d cases.

Template Parameters

StencilClass Stencil shall be given by a class

Parameters

prew	Function given in prewavelet basis
restrictions	2^d cases of prolongations (=0) and restrictions=1)
stencil

Here is the caller graph for this function:

◆ caseFunction_LocalStiffnessMatrix()

template<class Problem >

void MatrixVectorHomogen::caseFunction_LocalStiffnessMatrix	(	VectorSparseG &	prew,
		bool *	restrictions,
		Problem &	localStiffnessMatrixAllDepths
	)

private

Calculates one of the 2^d cases.

Parameters

prew	Function given in prewavelet basis
restrictions	2^d cases of prolongations (=0) and restrictions=1)
stencil

◆ casefunction_mpi()

template<class Problem >

void MatrixVectorHomogen::casefunction_mpi	(	VectorSparseG &	prew,
		VectorSparseG &	Ax,
		Problem &	stencilClass
	)

Distribute the cases with dynamic mpi_scheduling.

Template Parameters

Problem

Parameters

prew
Ax
stencilClass

◆ ConvertToPrewavelet2()

void MatrixVectorHomogen::ConvertToPrewavelet2	(	VectorSparseG &	Ax_nodal,
		VectorSparseG &	Ax,
		Depth &	T
	)

inlineprivate

Convert dualspace to prewavlet basis i.e. a(u,nodal_p) -> a(u,prewavelet_p)

Parameters

Ax_nodal	a(u,nodal_p)
Ax	a(u,prewavlet_p)
T	T(p) \leq T \forall p in D

Here is the caller graph for this function:

◆ intToBinary()

void MatrixVectorHomogen::intToBinary	(	int	num,
		bool	binary[]
	)

inlineprivate

Converts an integer to a binary and stores the resulting binary as a boolean.

Parameters

num
binary

Here is the caller graph for this function:

◆ master()

void MatrixVectorHomogen::master	(	int	num_workers,
		int	total_tasks
	)

private

This function is used for dynamic scheduling of the 2^d different cases of prolongations and restrictions. The master sends a case (bool restrictions) to the worker. The worker then starts calculation the required partial sum of the bilinear form.

Parameters

num_workers
total_tasks

◆ master_distribute()

void MatrixVectorHomogen::master_distribute ( )

private

This function is used for dynamic scheduling of the 2^d different cases of prolongations and restrictions. The master sends a case (bool restrictions) to the worker. The worker then starts calculation the required partial sum of the bilinear form.

◆ master_end_distribute()

void MatrixVectorHomogen::master_end_distribute ( )

private

This function is used for dynamic scheduling of the 2^d different cases of prolongations and restrictions. The master sends a case (bool restrictions) to the worker. The worker then starts calculation the required partial sum of the bilinear form.

Here is the caller graph for this function:

◆ master_start()

void MatrixVectorHomogen::master_start ( int num_workers )

private

This function is used for dynamic scheduling of the 2^d different cases of prolongations and restrictions. The master sends a case (bool restrictions) to the worker. The worker then starts calculation the required partial sum of the bilinear form.

Parameters

num_workers
total_tasks

Here is the caller graph for this function:

◆ multiplication()

template<class StencilClass >

void MatrixVectorHomogen::multiplication	(	VectorSparseG &	prew,
		VectorSparseG &	Ax,
		StencilClass &	stencilClass
	)

Calculates \( a(prew, \varphi_p) \quad \forall p \in D \) and stores it in Ax, with prew given in prewavelet basis.

Template Parameters

StencilClass

Parameters

prew
Ax
stencilClass

◆ prolongation1D_inplace()

void MatrixVectorHomogen::prolongation1D_inplace	(	VectorSparseG &	values,
		int &	t,
		int &	d
	)

inlineprivate

Prolongation operator --> prolongation on primal space

Parameters

values
t	depth
d	direction

Here is the caller graph for this function:

◆ restriction1D_inverted_inplace()

void MatrixVectorHomogen::restriction1D_inverted_inplace	(	const VectorSparseG &	values,
		int	t,
		int	d
	)

inlineprivate

Restriction operator -> Restriction of dual space

Parameters

values
t
d

Here is the caller graph for this function:

◆ worker()

template<class StencilClass >

void MatrixVectorHomogen::worker	(	VectorSparseG &	prew,
		VectorSparseG &	Ax,
		StencilClass &	stencilClass
	)

private

worker function for dynamic mpi scheduling of the cases

Template Parameters

StencilClass

Parameters

prew
Ax
stencilClass

Here is the caller graph for this function:

Member Data Documentation

◆ option_mpi

bool MatrixVectorHomogen::option_mpi = false

Distribute local stiffness matrices with MPI. (No speedup yet!)

◆ option_symmetry

bool MatrixVectorHomogen::option_symmetry = false

make use of symmetry of the bilinear form

◆ stencil_count

int MatrixVectorHomogen::stencil_count =0

counter for stencil evaluations, just needed for testing

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

/home/runner/work/LoAdSG/LoAdSG/library/source/MatrixVectorMultiplication/MatrixVectorHomogen.h
/home/runner/work/LoAdSG/LoAdSG/library/source/MatrixVectorMultiplication/MatrixVectorHomogen.cpp

Public Member Functions

Public Attributes

Private Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ MatrixVectorHomogen() [1/3]

◆ MatrixVectorHomogen() [2/3]

◆ MatrixVectorHomogen() [3/3]

Member Function Documentation

◆ applyStencilGlobal()

◆ applyStencilGlobal_MPI()

◆ applyStencilGlobal_MPI_2()

◆ applyStencilGlobal_MPI_3()

◆ applyStencilGlobal_symmetry()

◆ applyStencilLocal()

◆ calcNodal() [1/2]

◆ calcNodal() [2/2]

◆ caseFunction()

◆ caseFunction_LocalStiffnessMatrix()

◆ casefunction_mpi()

◆ ConvertToPrewavelet2()

◆ intToBinary()

◆ master()

◆ master_distribute()

◆ master_end_distribute()

◆ master_start()

◆ multiplication()

◆ prolongation1D_inplace()

◆ restriction1D_inverted_inplace()

◆ worker()

Member Data Documentation

◆ option_mpi

◆ option_symmetry

◆ stencil_count