#include <deal.II/lac/cuda_vector.h>

Public Types
using	value_type = Number

using	size_type = types::global_dof_index

using	real_type = typename numbers::NumberTraits<Number>::real_type

Public Member Functions
	Vector ()

	Vector (const Vector< Number > &V)

	Vector (Vector< Number > &&) noexcept=default

	Vector (const size_type n)

Vector &	operator= (const Vector< Number > &v)

Vector &	operator= (Vector< Number > &&v) noexcept=default

void	swap (Vector< Number > &v)

void	reinit (const size_type n, const bool omit_zeroing_entries=false)

void	reinit (const Vector< Number > &V, const bool omit_zeroing_entries=false)

void	import_elements (const ReadWriteVector< Number > &V, const VectorOperation::values operation, const std::shared_ptr< const Utilities::MPI::CommunicationPatternBase > &communication_pattern={})

void	import (const ReadWriteVector< Number > &V, VectorOperation::values operation, std::shared_ptr< const Utilities::MPI::CommunicationPatternBase > communication_pattern={})

Vector< Number > &	operator= (const Number s)

Vector< Number > &	operator*= (const Number factor)

Vector< Number > &	operator/= (const Number factor)

Vector< Number > &	operator+= (const Vector< Number > &V)

Vector< Number > &	operator-= (const Vector< Number > &V)

Number	operator* (const Vector< Number > &V) const

void	add (const Number a)

void	add (const Number a, const Vector< Number > &V)

void	add (const Number a, const Vector< Number > &V, const Number b, const Vector< Number > &W)

void	sadd (const Number s, const Number a, const Vector< Number > &V)

void	scale (const Vector< Number > &scaling_factors)

void	equ (const Number a, const Vector< Number > &V)

bool	all_zero () const

value_type	mean_value () const

real_type	l1_norm () const

real_type	l2_norm () const

real_type	norm_sqr () const

real_type	linfty_norm () const

Number	add_and_dot (const Number a, const Vector< Number > &V, const Vector< Number > &W)

Number *	get_values () const

size_type	size () const

::IndexSet	locally_owned_elements () const

void	print (std::ostream &out, const unsigned int precision=2, const bool scientific=true, const bool across=true) const

std::size_t	memory_consumption () const

Static Public Member Functions
static ::ExceptionBase &	ExcVectorTypeNotCompatible ()

Private Attributes
std::unique_ptr< Number[], void()(Number )>	val

size_type	n_elements

Detailed Description

template<typename Number>
class LinearAlgebra::CUDAWrappers::Vector< Number >

This class implements a vector using CUDA for use on Nvidia GPUs.

Note: Only float and double are supported.

See also: CUDAWrappers

Definition at line 55 of file cuda_vector.h.

Member Typedef Documentation

◆ value_type

template<typename Number >

using LinearAlgebra::CUDAWrappers::Vector< Number >::value_type = Number

Definition at line 58 of file cuda_vector.h.

◆ size_type

template<typename Number >

using LinearAlgebra::CUDAWrappers::Vector< Number >::size_type = types::global_dof_index

Definition at line 59 of file cuda_vector.h.

◆ real_type

template<typename Number >

using LinearAlgebra::CUDAWrappers::Vector< Number >::real_type = typename numbers::NumberTraits<Number>::real_type

Definition at line 60 of file cuda_vector.h.

Constructor & Destructor Documentation

◆ Vector() [1/4]

template<typename Number >

Vector< Number >::Vector ( )

Constructor. Create a vector of dimension zero.

Definition at line 40 of file cuda_vector.cc.

◆ Vector() [2/4]

template<typename Number >

Vector< Number >::Vector ( const Vector< Number > & V )

Copy constructor.

Definition at line 48 of file cuda_vector.cc.

◆ Vector() [3/4]

template<typename Number >

LinearAlgebra::CUDAWrappers::Vector< Number >::Vector ( Vector< Number > && )

defaultnoexcept

Move constructor.

◆ Vector() [4/4]

template<typename Number >

Vector< Number >::Vector ( const size_type n )

explicit

Constructor. Set dimension to n and initialize all elements with zero.

The constructor is made explicit to avoid accident like this: v=0;. Presumably, the user wants to set every elements of the vector to zero, but instead, what happens is this call: v=Vector<Number>(0);, i.e. the vector is replaced by one of length zero.

Definition at line 85 of file cuda_vector.cc.

Member Function Documentation

◆ operator=() [1/3]

template<typename Number >

Vector< Number > & Vector< Number >::operator= ( const Vector< Number > & v )

Copy assignment operator.

Definition at line 65 of file cuda_vector.cc.

◆ operator=() [2/3]

template<typename Number >

Vector & LinearAlgebra::CUDAWrappers::Vector< Number >::operator= ( Vector< Number > && v )

defaultnoexcept

Move assignment operator.

◆ swap()

template<typename Number >

void Vector< Number >::swap ( Vector< Number > & v )

inline

Swap the contents of this vector and the other vector v. One could do this operation with a temporary variable and copying over the data elements, but this function is significantly more efficient since it only swaps the pointers to the data of the two vectors and therefore does not need to allocate temporary storage and move data around.

This function is analogous to the swap function of all C++ standard containers. Also, there is a global function swap(u,v) that simply calls u.swap(v), again in analogy to standard functions.

Definition at line 405 of file cuda_vector.h.

◆ reinit() [1/2]

template<typename Number >

void Vector< Number >::reinit	(	const size_type	n,
		const bool	omit_zeroing_entries = false )

Reinit functionality. The flag omit_zeroing_entries determines whether the vector should be filled with zeros (false) or left in an undetermined state (true).

Definition at line 96 of file cuda_vector.cc.

◆ reinit() [2/2]

template<typename Number >

void Vector< Number >::reinit	(	const Vector< Number > &	V,
		const bool	omit_zeroing_entries = false )

Change the dimension to that of the vector V. The elements of V are not copied.

Definition at line 118 of file cuda_vector.cc.

◆ import_elements()

template<typename Number >

void Vector< Number >::import_elements	(	const ReadWriteVector< Number > &	V,
		const VectorOperation::values	operation,
		const std::shared_ptr< const Utilities::MPI::CommunicationPatternBase > &	communication_pattern = {} )

Import all the element from the input vector V. VectorOperation::values operation is used to decide if the elements int V should be added to the current vector or replace the current elements. The last parameter is not used. It is only used for distributed vectors. This is the function that should be used to copy a vector to the GPU.

Definition at line 128 of file cuda_vector.cc.

◆ import()

template<typename Number >

void LinearAlgebra::CUDAWrappers::Vector< Number >::import	(	const ReadWriteVector< Number > &	V,
		VectorOperation::values	operation,
		std::shared_ptr< const Utilities::MPI::CommunicationPatternBase >	communication_pattern = {} )

inline

Deprecated: Use import_elements() instead.

Definition at line 150 of file cuda_vector.h.

◆ operator=() [3/3]

template<typename Number >

Vector< Number > & Vector< Number >::operator= ( const Number s )

Sets all elements of the vector to the scalar s. This operation is only allowed if s is equal to zero.

Definition at line 174 of file cuda_vector.cc.

◆ operator*=()

template<typename Number >

Vector< Number > & Vector< Number >::operator*= ( const Number factor )

Multiply the entive vector by a fixed factor.

Definition at line 190 of file cuda_vector.cc.

◆ operator/=()

template<typename Number >

Vector< Number > & Vector< Number >::operator/= ( const Number factor )

Divide the entire vector by a fixed factor.

Definition at line 205 of file cuda_vector.cc.

◆ operator+=()

template<typename Number >

Vector< Number > & Vector< Number >::operator+= ( const Vector< Number > & V )

Add the vector V to the present one.

Definition at line 221 of file cuda_vector.cc.

◆ operator-=()

template<typename Number >

Vector< Number > & Vector< Number >::operator-= ( const Vector< Number > & V )

Subtract the vector V from the present one.

Definition at line 240 of file cuda_vector.cc.

◆ operator*()

template<typename Number >

Number Vector< Number >::operator* ( const Vector< Number > & V ) const

Return the scalar product of two vectors.

Definition at line 259 of file cuda_vector.cc.

◆ add() [1/3]

template<typename Number >

void Vector< Number >::add ( const Number a )

Add to all components. Note that a is a scalar not a vector.

Definition at line 296 of file cuda_vector.cc.

◆ add() [2/3]

template<typename Number >

void Vector< Number >::add	(	const Number	a,
		const Vector< Number > &	V )

Simple addition of a multiple of a vector, i.e. *this += a*V.

Definition at line 309 of file cuda_vector.cc.

◆ add() [3/3]

template<typename Number >

void Vector< Number >::add	(	const Number	a,
		const Vector< Number > &	V,
		const Number	b,
		const Vector< Number > &	W )

Multiple additions of scaled vectors, i.e. *this += a*V+b*W.

Definition at line 327 of file cuda_vector.cc.

◆ sadd()

template<typename Number >

void Vector< Number >::sadd	(	const Number	s,
		const Number	a,
		const Vector< Number > &	V )

Scaling and simple addition of a multiple of a vector, i.e. this = s(*this)+a*V

Definition at line 353 of file cuda_vector.cc.

◆ scale()

template<typename Number >

void Vector< Number >::scale ( const Vector< Number > & scaling_factors )

Scale each element of this vector by the corresponding element in the argument. This function is mostly meant to simulate multiplication (and immediate re-assignment) by a diagonal scaling matrix.

Definition at line 374 of file cuda_vector.cc.

◆ equ()

template<typename Number >

void Vector< Number >::equ	(	const Number	a,
		const Vector< Number > &	V )

Assignment *this = a*V.

Definition at line 392 of file cuda_vector.cc.

◆ all_zero()

template<typename Number >

bool Vector< Number >::all_zero ( ) const

Return whether the vector contains only elements with value zero.

Definition at line 413 of file cuda_vector.cc.

◆ mean_value()

template<typename Number >

Vector< Number >::value_type Vector< Number >::mean_value ( ) const

Return the mean value of all the entries of this vector.

Definition at line 422 of file cuda_vector.cc.

◆ l1_norm()

template<typename Number >

Vector< Number >::real_type Vector< Number >::l1_norm ( ) const

Return the l₁ norm of the vector (i.e., the sum of the absolute values of all entries among all processors).

Definition at line 453 of file cuda_vector.cc.

◆ l2_norm()

template<typename Number >

Vector< Number >::real_type Vector< Number >::l2_norm ( ) const

Return the l₂ norm of the vector (i.e., the square root of the sum of the square of all entries among all processors).

Definition at line 483 of file cuda_vector.cc.

◆ norm_sqr()

template<typename Number >

Vector< Number >::real_type Vector< Number >::norm_sqr ( ) const

Return the square of the \(l_2\)-norm.

Definition at line 492 of file cuda_vector.cc.

◆ linfty_norm()

template<typename Number >

Vector< Number >::real_type Vector< Number >::linfty_norm ( ) const

Return the maximum norm of the vector (i.e., the maximum absolute value among all entries and among all processors).

Definition at line 501 of file cuda_vector.cc.

◆ add_and_dot()

template<typename Number >

Number Vector< Number >::add_and_dot	(	const Number	a,
		const Vector< Number > &	V,
		const Vector< Number > &	W )

Perform a combined operation of a vector addition and a subsequent inner product, returning the value of the inner product. In other words, the result of this function is the same as if the user called

this->add(a, V);

return_value = *this * W;

LinearAlgebra::CUDAWrappers::Vector::add

void add(const Number a)

Definition cuda_vector.cc:296

The reason this function exists is that this operation involves less memory transfer than calling the two functions separately. This method only needs to load three vectors, this, V, W, whereas calling separate methods means to load the calling vector this twice. Since most vector operations are memory transfer limited, this reduces the time by 25% (or 50% if W equals this).

For complex-valued vectors, the scalar product in the second step is implemented as \(\left<v,w\right>=\sum_i v_i \bar{w_i}\).

Definition at line 531 of file cuda_vector.cc.

◆ get_values()

template<typename Number >

Number * Vector< Number >::get_values ( ) const

inline

Return the pointer to the underlying array. Ownership still resides with this class.

Definition at line 379 of file cuda_vector.h.

◆ size()

template<typename Number >

Vector< Number >::size_type Vector< Number >::size ( ) const

inline

Return the size of the vector.

Definition at line 388 of file cuda_vector.h.

◆ locally_owned_elements()

template<typename Number >

IndexSet Vector< Number >::locally_owned_elements ( ) const

inline

Return an index set that describe which elements of this vector are owned by the current processor, i.e. [0, size).

Definition at line 396 of file cuda_vector.h.

◆ print()

template<typename Number >

void Vector< Number >::print	(	std::ostream &	out,
		const unsigned int	precision = 2,
		const bool	scientific = true,
		const bool	across = true ) const

Print the vector to the output stream out.

Definition at line 566 of file cuda_vector.cc.

◆ memory_consumption()

template<typename Number >

std::size_t Vector< Number >::memory_consumption ( ) const

Return the memory consumption of this class in bytes.

Definition at line 602 of file cuda_vector.cc.

Member Data Documentation

◆ val

template<typename Number >

std::unique_ptr<Number[], void (*)(Number *)> LinearAlgebra::CUDAWrappers::Vector< Number >::val

private

Pointer to the array of elements of this vector.

Definition at line 346 of file cuda_vector.h.

◆ n_elements

template<typename Number >

size_type LinearAlgebra::CUDAWrappers::Vector< Number >::n_elements

private

Number of elements in the vector.

Definition at line 351 of file cuda_vector.h.

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

include/deal.II/lac/cuda_vector.h
source/lac/cuda_vector.cc

Public Types

Public Member Functions

Static Public Member Functions

Private Attributes

Detailed Description

Member Typedef Documentation

◆ value_type

◆ size_type

◆ real_type

Constructor & Destructor Documentation

◆ Vector() [1/4]

◆ Vector() [2/4]

◆ Vector() [3/4]

◆ Vector() [4/4]

Member Function Documentation

◆ operator=() [1/3]

◆ operator=() [2/3]

◆ swap()

◆ reinit() [1/2]

◆ reinit() [2/2]

◆ import_elements()

◆ import()

◆ operator=() [3/3]

◆ operator*=()

◆ operator/=()

◆ operator+=()

◆ operator-=()

◆ operator*()

◆ add() [1/3]

◆ add() [2/3]

◆ add() [3/3]

◆ sadd()

◆ scale()

◆ equ()

◆ all_zero()

◆ mean_value()

◆ l1_norm()

◆ l2_norm()

◆ norm_sqr()

◆ linfty_norm()

◆ add_and_dot()

◆ get_values()

◆ size()

◆ locally_owned_elements()

◆ print()

◆ memory_consumption()

Member Data Documentation

◆ val

◆ n_elements