Generic global assembler for operators on a gridview. More...

#include <dune/fufem/assemblers/dunefunctionsoperatorassembler.hh>

Public Member Functions
	DuneFunctionsOperatorAssembler (const TestBasis &tBasis, const AnsatzBasis &aBasis)
	create assembler for grid

template<class PatternBuilder >
void	assembleBulkPattern (PatternBuilder &patternBuilder) const

template<class PatternBuilder , class ElementPartition >
void	assembleBulkPattern (PatternBuilder &patternBuilder, const ElementPartition &elementPartition, std::size_t threadCount) const
	Thread parallel version of assembleBulkPattern.

template<class PatternBuilder >
void	assembleSkeletonPattern (PatternBuilder &patternBuilder) const

template<class Matrix , class LocalAssembler >
void	assembleBulkEntries (Matrix &&matrix, LocalAssembler &&localAssembler) const
	Assemble bulk integrals.

template<class Matrix , class LocalAssembler , class ElementPartition >
void	assembleBulkEntries (Matrix &&matrix, LocalAssembler &&localAssembler, const ElementPartition &elementPartition, std::size_t threadCount) const
	Thread parallel version of assembleBulkEntries.

template<class Matrix , class LocalAssembler , class LocalBoundaryAssembler >
void	assembleSkeletonEntries (Matrix &&matrix, LocalAssembler &&localAssembler, LocalBoundaryAssembler &&localBoundaryAssembler) const

template<class Matrix , class LocalAssembler >
void	assembleBulk (Matrix &&matrix, LocalAssembler &&localAssembler) const
	Create matrix pattern and assemble bulk integrals.

template<class Matrix , class LocalAssembler , class ElementPartition >
void	assembleBulk (Matrix &&matrix, LocalAssembler &&localAssembler, const ElementPartition &elementPartition, std::size_t threadCount) const
	Thread parallel version of assembleBulk.

Protected Attributes
const TestBasis &	testBasis_

const AnsatzBasis &	ansatzBasis_

Detailed Description

template<class TestBasis, class AnsatzBasis>
class Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >

Generic global assembler for operators on a gridview.

Deprecated:: This class is deprecated and will be removed after 2.11. Use the global assemblers from the dune-assember module instead.

Constructor & Destructor Documentation

◆ DuneFunctionsOperatorAssembler()

template<class TestBasis , class AnsatzBasis >

Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::DuneFunctionsOperatorAssembler	(	const TestBasis &	tBasis,
		const AnsatzBasis &	aBasis
	)

inline

create assembler for grid

Member Function Documentation

◆ assembleBulk() [1/2]

template<class TestBasis , class AnsatzBasis >

template<class Matrix , class LocalAssembler >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulk	(	Matrix &&	matrix,
		LocalAssembler &&	localAssembler
	)		const

inline

Create matrix pattern and assemble bulk integrals.

This will first create a matrix pattern using assembleBulkPattern() and then compute the entries using assembleBulkEntries(). If passed matrix does not satisfy the MatrixBackend concept, it is wrapped using istlMatrixBackend().

Parameters

matrix	The matrix to be assembled.
localAssembler	The local element assembler

◆ assembleBulk() [2/2]

template<class TestBasis , class AnsatzBasis >

template<class Matrix , class LocalAssembler , class ElementPartition >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulk	(	Matrix &&	matrix,
		LocalAssembler &&	localAssembler,
		const ElementPartition &	elementPartition,
		std::size_t	threadCount
	)		const

inline

Thread parallel version of assembleBulk.

This will first create a matrix pattern using assembleBulkPattern() and then compute the entries using assembleBulkEntries(). If passed matrix does not satisfy the MatrixBackend concept, it is wrapped using istlMatrixBackend().

Given a colored elementPartition of the grid view associated to the bases, this assembles the problem by processing each color from the partition with threadCount parallel threads. The colored elementPartition should be a range of ranges containing one random-access element range per color. The elements associated to each color are distributed (approximately) uniformly among the threads. After one color is processed, the threads are synchronized before proceeding with the next color.

This requires that the colorings represent a disjoint partition of all elements and that each color is overlap-free in the following sense: Given two elements e1 and e2, from a single color, then the sets of basis functions associated with e1 and e2 are disjoint.

This will use one copy of the local assembler per thread. Hence it must be guaranteed that this is thread-safe. E.g. any mutable data like internal caches must not be shared.

Parameters

matrix	The matrix to be assembled.
localAssembler	The local element assembler
elementPartition	A partition of the elements in the grid view as e.g. provided by coloredGridViewPartition()
threadCount	Number of threads to use.

◆ assembleBulkEntries() [1/2]

template<class TestBasis , class AnsatzBasis >

template<class Matrix , class LocalAssembler >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulkEntries	(	Matrix &&	matrix,
		LocalAssembler &&	localAssembler
	)		const

inline

Assemble bulk integrals.

If passed matrix does not satisfy the MatrixBackend concept, it is wrapped using istlMatrixBackend().

Parameters

matrix	The matrix to be assembled.
localAssembler	The local element assembler

◆ assembleBulkEntries() [2/2]

template<class TestBasis , class AnsatzBasis >

template<class Matrix , class LocalAssembler , class ElementPartition >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulkEntries	(	Matrix &&	matrix,
		LocalAssembler &&	localAssembler,
		const ElementPartition &	elementPartition,
		std::size_t	threadCount
	)		const

inline

Thread parallel version of assembleBulkEntries.

If passed matrix does not satisfy the MatrixBackend concept, it is wrapped using istlMatrixBackend().

Cf. the documentation of the respective thread parallel assembleBulk() method for details and constraints of the used parallel approach.

Parameters

matrix	The matrix to be assembled.
localAssembler	The local element assembler
elementPartition	A partition of the elements in the grid view as e.g. provided by coloredGridViewPartition()
threadCount	Number of threads to use.

◆ assembleBulkPattern() [1/2]

template<class TestBasis , class AnsatzBasis >

template<class PatternBuilder >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulkPattern ( PatternBuilder & patternBuilder ) const

inline

◆ assembleBulkPattern() [2/2]

template<class TestBasis , class AnsatzBasis >

template<class PatternBuilder , class ElementPartition >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleBulkPattern	(	PatternBuilder &	patternBuilder,
		const ElementPartition &	elementPartition,
		std::size_t	threadCount
	)		const

inline

Thread parallel version of assembleBulkPattern.

Cf. the documentation of the respective thread parallel assembleBulk() method for details and constraints of the used parallel approach.

Parameters

patternBuilder	The object to store the pattern
elementPartition	A partition of the elements in the grid view as e.g. provided by coloredGridViewPartition()
threadCount	Number of threads to use.

◆ assembleSkeletonEntries()

template<class TestBasis , class AnsatzBasis >

template<class Matrix , class LocalAssembler , class LocalBoundaryAssembler >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleSkeletonEntries	(	Matrix &&	matrix,
		LocalAssembler &&	localAssembler,
		LocalBoundaryAssembler &&	localBoundaryAssembler
	)		const

inline

◆ assembleSkeletonPattern()

template<class TestBasis , class AnsatzBasis >

template<class PatternBuilder >

void Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::assembleSkeletonPattern ( PatternBuilder & patternBuilder ) const

inline

Member Data Documentation

◆ ansatzBasis_

template<class TestBasis , class AnsatzBasis >

const AnsatzBasis& Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::ansatzBasis_

protected

◆ testBasis_

template<class TestBasis , class AnsatzBasis >

const TestBasis& Dune::Fufem::DuneFunctionsOperatorAssembler< TestBasis, AnsatzBasis >::testBasis_

protected

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

dunefunctionsoperatorassembler.hh

Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ DuneFunctionsOperatorAssembler()

Member Function Documentation

◆ assembleBulk() [1/2]

◆ assembleBulk() [2/2]

◆ assembleBulkEntries() [1/2]

◆ assembleBulkEntries() [2/2]

◆ assembleBulkPattern() [1/2]

◆ assembleBulkPattern() [2/2]

◆ assembleSkeletonEntries()

◆ assembleSkeletonPattern()

Member Data Documentation

◆ ansatzBasis_

◆ testBasis_