tarch::la::DynamicMatrix Class Reference

My standard matrix is a matrix where the size is fixed at compile time. More...

#include <DynamicMatrix.h>

Public Member Functions
	DynamicMatrix (int rows, int cols)
	Create empty matrix.
	~DynamicMatrix ()
	Free the array on the heap.
	DynamicMatrix (int rows, int cols, std::initializer_list< std::initializer_list< double > > values)
	DynamicMatrix (const DynamicMatrix &)
	DynamicMatrix (DynamicMatrix &&)
	DynamicMatrix (const DynamicMatrix &lhs, const DynamicMatrix &rhs, bool innerProduct)
	DynamicMatrix (int rows, int cols, double *data)
DynamicMatrix &	operator= (const DynamicMatrix &)=delete
DynamicMatrix &	operator= (std::initializer_list< std::initializer_list< double > > values)
	This template allows you to write stuff like.
template<int Cols>
bool	operator== (double values[][Cols]) const
bool	operator== (const DynamicMatrix &matrix) const
std::string	toString (bool addLineBreaks=false) const
double *	data ()
const double *	data () const
double &	operator() (int row, int col)
double	operator() (int row, int col) const
void	multiply (double *result, double *x)
void	multiplyBySmallMatrix (double *result, const DynamicMatrix &matrix) const
void	batchedMultiplyAoS (double *__restrict__ result, const double *__restrict__ x, int batchCount, int resultSize, int firstRow)
	This operation assumes that x holds a whole batch of vectors in AoS format.
void	batchedMultiplyAoS (double *__restrict__ result, const double *__restrict__ x, int batchCount)
	Wrapper around other batch operation.
void	replicateRows (int blockSize, int numberOfReplications, int shiftAfterEveryReplication, bool extendColumnsToAccommodateShifts)
	Split the matrix into blocks of rows of size blockSize.
void	replicateCols (int blockSize, int numberOfReplications, int shiftAfterEveryReplication, bool extendColumnsToAccommodateShifts)
void	insertEmptyColumns (int number, int where, int repeatEveryKColumns=0)
	Insert zero columns.
void	insertEmptyRows (int number, int where, int repeatEveryKColumns=0)
void	shiftRowsDown (int shift, bool wrap=false)
	Shift the rows to the right.
void	shiftColumnsRight (int shift, bool wrap=false)
void	removeColumn (int number)
void	scale (double value)
	Scale all entries.
int	rows () const
int	cols () const

Static Public Member Functions
static std::string	vectorToString (double *values, int entries, bool addLineBreaks=false)
	I often need this in combination with the toString() operation above.
static tarch::la::DynamicMatrix	id (int rows)
	Create (square) identify matrix with rows rows and column.

Private Member Functions
int	serialise (int row, int col) const

Static Private Member Functions
static int	serialise (int row, int col, int Rows, int Cols)

Private Attributes
int	_cols
int	_rows
double *	_m

Detailed Description

My standard matrix is a matrix where the size is fixed at compile time.

This is a matrix class where the size is fixed at runtime.

Definition at line 20 of file DynamicMatrix.h.

Constructor & Destructor Documentation

DynamicMatrix() [1/6]

tarch::la::DynamicMatrix::DynamicMatrix	(	int	rows,
		int	cols
	)

Create empty matrix.

~DynamicMatrix()

tarch::la::DynamicMatrix::~DynamicMatrix

(

)

Free the array on the heap.

If the matrix has been moved before (due to return value optimisation, e.g.) _m is equal to nullptr and we should not delete it anymore. Otherwise, _m always has to be freed.

DynamicMatrix() [2/6]

tarch::la::DynamicMatrix::DynamicMatrix	(	int	rows,
		int	cols,
		std::initializer_list< std::initializer_list< double > >	values
	)

DynamicMatrix() [3/6]

tarch::la::DynamicMatrix::DynamicMatrix

(

const DynamicMatrix &

)

DynamicMatrix() [4/6]

tarch::la::DynamicMatrix::DynamicMatrix

(

DynamicMatrix &&

)

DynamicMatrix() [5/6]

tarch::la::DynamicMatrix::DynamicMatrix	(	const DynamicMatrix &	lhs,
		const DynamicMatrix &	rhs,
		bool	innerProduct
	)

Parameters

innerProduct

If this flag is set, I use the inner product between the two matrices. If it is not set, I use the generalised outer product, i.e. the Kronecker product to construct the new matrix.

DynamicMatrix() [6/6]

tarch::la::DynamicMatrix::DynamicMatrix	(	int	rows,
		int	cols,
		double *	data
	)

Member Function Documentation

batchedMultiplyAoS() [1/2]

void tarch::la::DynamicMatrix::batchedMultiplyAoS	(	double *__restrict__	result,
		const double *__restrict__	x,
		int	batchCount
	)

Wrapper around other batch operation.

The whole matrix is used.

batchedMultiplyAoS() [2/2]

void tarch::la::DynamicMatrix::batchedMultiplyAoS	(	double *__restrict__	result,
		const double *__restrict__	x,
		int	batchCount,
		int	resultSize,
		int	firstRow
	)

This operation assumes that x holds a whole batch of vectors in AoS format.

If batchSize is 5, then x points to five vectors of length _rows. The five vectors are store as AoS, i.e. x[0] is the first entry of the first vector, x[1] the first entry of the second vector, and so forth. The multiplication applies the matrix to all vectors stored in x. It assumes that result also holds a sequence of vectors.

Parameters

batchSize	Number of elements within x/result to which the matrix has to be applied in one rush.
resultSize	Number of entries in vector result. Please note that this number will be multiplied with batchSize, i.e. it is the entries per batch entry.
firstRow	First row to pick from matrix. If the resultSize is equal to the number of rows of the matrix, then this entry should be 0. Otherwise, you can pick only a few rows from the matrix through this offset. The fact that I can use a subset of the matrix requires me to pass in resultSize explicitly.

cols()

int tarch::la::DynamicMatrix::cols

(

)

const

data() [1/2]

double * tarch::la::DynamicMatrix::data

(

)

Definition at line 130 of file DynamicMatrix.h.

References _m.

Referenced by peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__generate_dastgen_input_file(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__get_file_name(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__get_full_qualified_file_name(), exaparty.api.particle.Particle.Particle::__init__(), exaparty.api.particle.Particle.Particle::_add_dependency_checks(), exaparty.api.particle.Particle.Particle::_dependency_checks_modify_steps(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_boolean_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_boolean_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_double_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_double_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_enum_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_integer_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_integer_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_peano_double_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_peano_integer_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_string_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_user_defined_attributes(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::_get_dictionary_for_output(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::construct_output(), peano4.api.datamodel.DynamicArrayOverPrimitivesToStdVector.DynamicArrayOverPrimitivesToStdVector::construct_output(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::construct_output(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::construct_output(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::construct_output(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::get_header_file_include(), peano4.api.datamodel.DynamicArrayOverPrimitivesToStdVector.DynamicArrayOverPrimitivesToStdVector::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::get_header_file_include(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::get_stack_container(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::get_stack_container(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::get_stack_container(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::get_stack_container(), peano4.api.datamodel.DaStGen2.DaStGen2::hosts_smart_pointer_attribute(), exaparty.api.particle.Particle.Particle::readme_descriptor(), exaparty.api.particle.SPHLeapfrogFixedSearchRadius.SPHLeapfrogFixedSearchRadius::set_parameters(), exaparty.api.particle.SPHParticle.SPHParticle::set_parameters(), and exaparty.api.particle.tests.testLeapfrogFixedTimeStepSize.testLeapfrogFixedTimeStepSize::set_parameters().

Here is the caller graph for this function:

data() [2/2]

const double * tarch::la::DynamicMatrix::data

(

)

const

Definition at line 131 of file DynamicMatrix.h.

References _m.

Referenced by peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__generate_dastgen_input_file(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__get_file_name(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::__get_full_qualified_file_name(), exaparty.api.particle.Particle.Particle::__init__(), exaparty.api.particle.Particle.Particle::_add_dependency_checks(), exaparty.api.particle.Particle.Particle::_dependency_checks_modify_steps(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_boolean_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_boolean_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_double_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_double_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_enum_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_integer_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_integer_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_peano_double_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_peano_integer_array_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_string_attributes(), exaparty.api.particle.tests.DastgenTestDummyParticle.DastgenTestDummyParticle::_generate_user_defined_attributes(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::_get_dictionary_for_output(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::construct_output(), peano4.api.datamodel.DynamicArrayOverPrimitivesToStdVector.DynamicArrayOverPrimitivesToStdVector::construct_output(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::construct_output(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::construct_output(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::construct_output(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::get_header_file_include(), peano4.api.datamodel.DynamicArrayOverPrimitivesToStdVector.DynamicArrayOverPrimitivesToStdVector::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::get_header_file_include(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::get_header_file_include(), peano4.api.datamodel.DaStGenToLegacyTool.DaStGenToLegacyTool::get_stack_container(), peano4.api.datamodel.PatchToDoubleArray.PatchToDoubleArray::get_stack_container(), peano4.api.datamodel.PatchToDoubleArrayOnHeap.PatchToDoubleArrayOnHeap::get_stack_container(), peano4.api.datamodel.PatchToDoubleArrayWithSmartPointer.PatchToDoubleArrayWithSmartPointer::get_stack_container(), peano4.api.datamodel.DaStGen2.DaStGen2::hosts_smart_pointer_attribute(), exaparty.api.particle.Particle.Particle::readme_descriptor(), exaparty.api.particle.SPHLeapfrogFixedSearchRadius.SPHLeapfrogFixedSearchRadius::set_parameters(), exaparty.api.particle.SPHParticle.SPHParticle::set_parameters(), and exaparty.api.particle.tests.testLeapfrogFixedTimeStepSize.testLeapfrogFixedTimeStepSize::set_parameters().

Here is the caller graph for this function:

id()

static tarch::la::DynamicMatrix tarch::la::DynamicMatrix::id ( int  rows )

static

Create (square) identify matrix with rows rows and column.

Referenced by exahype2.api.dslhype.SyntaxTree.Name::__eq__(), exahype2.api.dslhype.SyntaxTree.Name::__gt__(), exahype2.api.dslhype.SyntaxTree.Name::index(), exahype2.api.dslhype.SyntaxTree.Name::index(), exahype2.api.dslhype.SyntaxTree.Name::multidimensional_index(), exahype2.api.dslhype.SyntaxTree.DataBlock::offset(), exahype2.api.dslhype.SyntaxTree.FaceDataBlock::offset(), exahype2.api.dslhype.SyntaxTree.Argument::print_cpp(), exahype2.api.dslhype.SyntaxTree.Name::print_cpp(), exahype2.api.dslhype.SyntaxTree.DataBlock::print_cpp(), exahype2.api.dslhype.SyntaxTree.FunctionDefinition::print_definition_with_timer(), exahype2.api.dslhype.SyntaxTree.Name::print_mlir(), exahype2.api.dslhype.SyntaxTree.DataBlock::print_mlir(), exahype2.api.dslhype.SyntaxTree.Argument::print_omp(), exahype2.api.dslhype.SyntaxTree.Name::print_omp(), exahype2.api.dslhype.SyntaxTree.DataBlock::print_omp(), exahype2.api.dslhype.SyntaxTree.Argument::print_sycl(), exahype2.api.dslhype.SyntaxTree.Name::print_sycl(), exahype2.api.dslhype.SyntaxTree.DataBlock::print_sycl(), exahype2.api.dslhype.SyntaxTree.Name::print_tree(), and exahype2.api.dslhype.SyntaxTree.DataBlock::print_tree().

Here is the caller graph for this function:

insertEmptyColumns()

void tarch::la::DynamicMatrix::insertEmptyColumns	(	int	number,
		int	where,
		int	repeatEveryKColumns = 0
	)

Insert zero columns.

Parameters

number	How many columns shall be inserted. Has to be at least one.
where	Number of the first column to be inserted. Put in 0 to insert one to the left of the matrix.
repeatEveryKColumns	If this one is zero, we insert only once. If it is bigger then zero, we insert number columns every repeatEveryKColumns starting from where. So if you pass in one, then every other column, we'll insert number of columns.

insertEmptyRows()

void tarch::la::DynamicMatrix::insertEmptyRows	(	int	number,
		int	where,
		int	repeatEveryKColumns = 0
	)

multiply()

void tarch::la::DynamicMatrix::multiply	(	double *	result,
		double *	x
	)

multiplyBySmallMatrix()

void tarch::la::DynamicMatrix::multiplyBySmallMatrix	(	double *	result,
		const DynamicMatrix &	matrix
	)		const

operator()() [1/2]

double & tarch::la::DynamicMatrix::operator()	(	int	row,
		int	col
	)

operator()() [2/2]

double tarch::la::DynamicMatrix::operator()	(	int	row,
		int	col
	)		const

operator=() [1/2]

DynamicMatrix & tarch::la::DynamicMatrix::operator= ( const DynamicMatrix &  )

delete

operator=() [2/2]

DynamicMatrix & tarch::la::DynamicMatrix::operator=

(

std::initializer_list< std::initializer_list< double > >

values

)

This template allows you to write stuff like.

double P[Rows][Cols] = {
  {3.0/3.0, 0.0,    0.0/3.0, 0.0,     0.0/3.0, 0.0,    0.0/3.0, 0.0},
  {    0.0, 0.0,        0.0, 0.0,         0.0, 0.0,        0.0, 0.0},
  {3.0/3.0, 0.0,    0.0/3.0, 0.0,     0.0/3.0, 0.0,    0.0/3.0, 0.0},
  {    0.0, 0.0,        0.0, 0.0,         0.0, 0.0,        0.0, 0.0},
  ...

tarch::la::DynamicMatrix myP(Rows,Cols);
myP = P;

   

operator==() [1/2]

bool tarch::la::DynamicMatrix::operator==

(

const DynamicMatrix &

matrix

)

const

operator==() [2/2]

template<int Cols>

bool tarch::la::DynamicMatrix::operator==

(

double

values[][Cols]

)

const

Definition at line 109 of file DynamicMatrix.h.

References _cols, _m, _rows, tarch::la::col(), tarch::la::equals(), tarch::la::row(), and serialise().

Here is the call graph for this function:

removeColumn()

void tarch::la::DynamicMatrix::removeColumn

(

int

number

)

replicateCols()

void tarch::la::DynamicMatrix::replicateCols	(	int	blockSize,
		int	numberOfReplications,
		int	shiftAfterEveryReplication,
		bool	extendColumnsToAccommodateShifts
	)

replicateRows()

void tarch::la::DynamicMatrix::replicateRows	(	int	blockSize,
		int	numberOfReplications,
		int	shiftAfterEveryReplication,
		bool	extendColumnsToAccommodateShifts
	)

Split the matrix into blocks of rows of size blockSize.

The number of rows now is increased by a factor of numberOfReplications. Each block is replicated numberOfReplications in the final output. For the replicated blocks, we shift each column by shiftAfterEveryReplication. So this entry can be zero if you want no shifts. Due to the shifts, it might happen that we access additional columns. If extendColumnsToAccommodateShifts is set, the operation appends columns as required. If extendColumnsToAccommodateShifts, it simply truncates the shifts.

rows()

int tarch::la::DynamicMatrix::rows

(

)

const

scale()

void tarch::la::DynamicMatrix::scale

(

double

value

)

Scale all entries.

serialise() [1/2]

int tarch::la::DynamicMatrix::serialise ( int  row, int  col  ) const

private

Definition at line 35 of file DynamicMatrix.h.

References _cols, _rows, tarch::la::col(), tarch::la::row(), and serialise().

Here is the call graph for this function:

serialise() [2/2]

static int tarch::la::DynamicMatrix::serialise ( int  row, int  col, int  Rows, int  Cols  )

staticprivate

Definition at line 26 of file DynamicMatrix.h.

References tarch::la::col(), and tarch::la::row().

Referenced by operator==(), and serialise().

Here is the call graph for this function:

Here is the caller graph for this function:

shiftColumnsRight()

void tarch::la::DynamicMatrix::shiftColumnsRight	(	int	shift,
		bool	wrap = false
	)

shiftRowsDown()

void tarch::la::DynamicMatrix::shiftRowsDown	(	int	shift,
		bool	wrap = false
	)

Shift the rows to the right.

toString()

std::string tarch::la::DynamicMatrix::toString

(

bool

addLineBreaks = false

)

const

vectorToString()

static std::string tarch::la::DynamicMatrix::vectorToString ( double *  values, int  entries, bool  addLineBreaks = false  )

static

I often need this in combination with the toString() operation above.

Field Documentation

_cols

int tarch::la::DynamicMatrix::_cols

private

Definition at line 22 of file DynamicMatrix.h.

Referenced by operator==(), and serialise().

_m

double* tarch::la::DynamicMatrix::_m

private

Definition at line 24 of file DynamicMatrix.h.

Referenced by data(), data(), and operator==().

_rows

int tarch::la::DynamicMatrix::_rows

private

Definition at line 23 of file DynamicMatrix.h.

Referenced by operator==(), and serialise().

---

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

• src/tarch/la/DynamicMatrix.h