The Ginkgo namespace. More...

Namespaces
	accessor
	The accessor namespace.

	factorization
	The Factorization namespace.

	log
	The logger namespace .

	matrix
	The matrix namespace.

	multigrid
	The multigrid components namespace.

	name_demangling
	The name demangling namespace.

	preconditioner
	The Preconditioner namespace.

	reorder
	The Reorder namespace.

	solver
	The ginkgo Solve namespace.

	stop
	The Stopping criterion namespace.

	syn
	The Synthesizer namespace.

	xstd
	The namespace for functionalities after C++14 standard.

Classes
class	AbsoluteComputable
	The AbsoluteComputable is an interface that allows to get the component wise absolute of a LinOp. More...

class	AbstractFactory
	The AbstractFactory is a generic interface template that enables easy implementation of the abstract factory design pattern. More...

class	AllocationError
	AllocationError is thrown if a memory allocation fails. More...

class	Allocator
	Provides generic allocation and deallocation functionality to be used by an Executor. More...

class	amd_device
	amd_device handles the number of executor on Amd devices and have the corresponding recursive_mutex. More...

struct	are_all_integral
	Evaluates if all template arguments Args fulfill std::is_integral. More...

struct	are_all_integral< First, Args... >

class	array
	An array is a container which encapsulates fixed-sized arrays, stored on the Executor tied to the array. More...

class	BadDimension
	BadDimension is thrown if an operation is being applied to a LinOp with bad dimensions. More...

struct	batch_dim
	A type representing the dimensions of a multidimensional batch object. More...

class	BlockOperator
	A BlockOperator represents a linear operator that is partitioned into multiple blocks. More...

class	BlockSizeError
	Error that denotes issues between block sizes and matrix dimensions. More...

class	Combination
	The Combination class can be used to construct a linear combination of multiple linear operators `c1 * op1 + c2 * op2 + ...` More...

class	Composition
	The Composition class can be used to compose linear operators `op1, op2, ..., opn` and obtain the operator `op1 * op2 * ...` More...

class	ConvertibleTo
	ConvertibleTo interface is used to mark that the implementer can be converted to the object of ResultType. More...

class	CpuAllocator
	Allocator using new/delete. More...

class	CpuAllocatorBase
	Implement this interface to provide an allocator for OmpExecutor or ReferenceExecutor. More...

class	CpuTimer
	A timer using std::chrono::steady_clock for timing. More...

struct	cpx_real_type
	Access the underlying real type of a complex number. More...

struct	cpx_real_type< std::complex< T > >
	Specialization for complex types.

class	CublasError
	CublasError is thrown when a cuBLAS routine throws a non-zero error code. More...

class	cuda_stream
	An RAII wrapper for a custom CUDA stream. More...

class	CudaAllocator
	Allocator using cudaMalloc. More...

class	CudaAllocatorBase
	Implement this interface to provide an allocator for CudaExecutor. More...

class	CudaAsyncAllocator

class	CudaError
	CudaError is thrown when a CUDA routine throws a non-zero error code. More...

class	CudaExecutor
	This is the Executor subclass which represents the CUDA device. More...

class	CudaHostAllocator

class	CudaTimer
	A timer using events for timing on a CudaExecutor. More...

class	CudaUnifiedAllocator

class	CufftError
	CufftError is thrown when a cuFFT routine throws a non-zero error code. More...

class	CurandError
	CurandError is thrown when a cuRAND routine throws a non-zero error code. More...

class	CusparseError
	CusparseError is thrown when a cuSPARSE routine throws a non-zero error code. More...

struct	default_converter
	Used to convert objects of type `S` to objects of type `R` using static_cast. More...

class	deferred_factory_parameter
	Represents a factory parameter of factory type that can either initialized by a pre-existing factory or by passing in a factory_parameters object whose `.on(exec)` will be called to instantiate a factory. More...

class	device_matrix_data
	This type is a device-side equivalent to matrix_data. More...

class	DiagonalExtractable
	The diagonal of a LinOp implementing this interface can be extracted. More...

class	DiagonalLinOpExtractable
	The diagonal of a LinOp can be extracted. More...

struct	dim
	A type representing the dimensions of a multidimensional object. More...

struct	dim< 1u, DimensionType >

class	DimensionMismatch
	DimensionMismatch is thrown if an operation is being applied to LinOps of incompatible size. More...

class	DpcppExecutor
	This is the Executor subclass which represents a DPC++ enhanced device. More...

class	DpcppTimer
	A timer using kernels for timing on a DpcppExecutor in profiling mode. More...

class	enable_parameters_type
	The enable_parameters_type mixin is used to create a base implementation of the factory parameters structure. More...

class	EnableAbsoluteComputation
	The EnableAbsoluteComputation mixin provides the default implementations of `compute_absolute_linop` and the absolute interface. More...

class	EnableAbstractPolymorphicObject
	This mixin inherits from (a subclass of) PolymorphicObject and provides a base implementation of a new abstract object. More...

class	EnableCreateMethod
	This mixin implements a static `create()` method on `ConcreteType` that dynamically allocates the memory, uses the passed-in arguments to construct the object, and returns an std::unique_ptr to such an object. More...

class	EnableDefaultFactory
	This mixin provides a default implementation of a concrete factory. More...

class	EnableLinOp
	The EnableLinOp mixin can be used to provide sensible default implementations of the majority of the LinOp and PolymorphicObject interface. More...

class	EnablePolymorphicAssignment
	This mixin is used to enable a default PolymorphicObject::copy_from() implementation for objects that have implemented conversions between them. More...

class	EnablePolymorphicObject
	This mixin inherits from (a subclass of) PolymorphicObject and provides a base implementation of a new concrete polymorphic object. More...

struct	err

class	Error
	The Error class is used to report exceptional behaviour in library functions. More...

class	Executor
	The first step in using the Ginkgo library consists of creating an executor. More...

class	executor_deleter
	This is a deleter that uses an executor's `free` method to deallocate the data. More...

class	executor_deleter< T[]>

class	hip_stream
	An RAII wrapper for a custom HIP stream. More...

class	HipAllocator

class	HipAllocatorBase
	Implement this interface to provide an allocator for HipExecutor. More...

class	HipAsyncAllocator

class	HipblasError
	HipblasError is thrown when a hipBLAS routine throws a non-zero error code. More...

class	HipError
	HipError is thrown when a HIP routine throws a non-zero error code. More...

class	HipExecutor
	This is the Executor subclass which represents the HIP enhanced device. More...

class	HipfftError
	HipfftError is thrown when a hipFFT routine throws a non-zero error code. More...

class	HipHostAllocator

class	HiprandError
	HiprandError is thrown when a hipRAND routine throws a non-zero error code. More...

class	HipsparseError
	HipsparseError is thrown when a hipSPARSE routine throws a non-zero error code. More...

class	HipTimer
	A timer using events for timing on a HipExecutor. More...

class	HipUnifiedAllocator

class	index_set
	An index set class represents an ordered set of intervals. More...

class	InvalidStateError
	Exception thrown if an object is in an invalid state. More...

class	KernelNotFound
	KernelNotFound is thrown if Ginkgo cannot find a kernel which satisfies the criteria imposed by the input arguments. More...

class	LinOp

class	LinOpFactory
	A LinOpFactory represents a higher order mapping which transforms one linear operator into another. More...

class	machine_topology
	The machine topology class represents the hierarchical topology of a machine, including NUMA nodes, cores and PCI Devices. More...

class	matrix_assembly_data
	This structure is used as an intermediate type to assemble a sparse matrix. More...

struct	matrix_data
	This structure is used as an intermediate data type to store a sparse matrix. More...

struct	matrix_data_entry
	Type used to store nonzeros. More...

class	MetisError
	MetisError is thrown when METIS routine throws an error code. More...

class	MpiError
	MpiError is thrown when a MPI routine throws a non-zero error code. More...

class	NotCompiled
	NotCompiled is thrown when attempting to call an operation which is a part of a module that was not compiled on the system. More...

class	NotImplemented
	NotImplemented is thrown in case an operation has not yet been implemented (but will be implemented in the future). More...

class	NotSupported
	NotSupported is thrown in case it is not possible to perform the requested operation on the given object type. More...

class	null_deleter
	This is a deleter that does not delete the object. More...

class	null_deleter< T[]>

class	nvidia_device
	nvidia_device handles the number of executor on Nvidia devices and have the corresponding recursive_mutex. More...

class	OmpExecutor
	This is the Executor subclass which represents the OpenMP device (typically CPU). More...

class	Operation
	Operations can be used to define functionalities whose implementations differ among devices. More...

class	OutOfBoundsError
	OutOfBoundsError is thrown if a memory access is detected to be out-of-bounds. More...

class	OverflowError
	OverflowError is thrown when an index calculation for storage requirements overflows. More...

class	Permutable
	Linear operators which support permutation should implement the Permutable interface. More...

class	Perturbation
	The Perturbation class can be used to construct a LinOp to represent the operation `(identity + scalar * basis * projector)`. More...

class	PolymorphicObject
	A PolymorphicObject is the abstract base for all "heavy" objects in Ginkgo that behave polymorphically. More...

class	precision_reduction
	This class is used to encode storage precisions of low precision algorithms. More...

class	Preconditionable
	A LinOp implementing this interface can be preconditioned. More...

class	ptr_param
	This class is used for function parameters in the place of raw pointers. More...

class	range
	A range is a multidimensional view of the memory. More...

class	ReadableFromMatrixData
	A LinOp implementing this interface can read its data from a matrix_data structure. More...

class	ReferenceExecutor
	This is a specialization of the OmpExecutor, which runs the reference implementations of the kernels used for debugging purposes. More...

class	ScaledIdentityAddable
	Adds the operation M <- a I + b M for matrix M, identity operator I and scalars a and b, where M is the calling object. More...

class	scoped_device_id_guard
	This move-only class uses RAII to set the device id within a scoped block, if necessary. More...

struct	segmented_array
	A minimal interface for a segmented array. More...

struct	span
	A span is a lightweight structure used to create sub-ranges from other ranges. More...

class	stopping_status
	This class is used to keep track of the stopping status of one vector. More...

class	StreamError
	StreamError is thrown if accessing a stream failed. More...

class	time_point
	An opaque wrapper for a time point generated by a timer. More...

class	Timer
	Represents a generic timer that can be used to record time points and measure time differences on host or device streams. More...

class	Transposable
	Linear operators which support transposition should implement the Transposable interface. More...

class	truncated

class	UnsupportedMatrixProperty
	Exception throws if a matrix does not have a property required by a numerical method. More...

class	UseComposition
	The UseComposition class can be used to store the composition information in LinOp. More...

class	ValueMismatch
	ValueMismatch is thrown if two values are not equal. More...

struct	version
	This structure is used to represent versions of various Ginkgo modules. More...

class	version_info
	Ginkgo uses version numbers to label new features and to communicate backward compatibility guarantees: More...

class	WritableToMatrixData
	A LinOp implementing this interface can write its data to a matrix_data structure. More...

Typedefs
template<typename ValueType >
using	Array = array< ValueType >

template<typename ConcreteFactory , typename ConcreteLinOp , typename ParametersType , typename PolymorphicBase = LinOpFactory>
using	EnableDefaultLinOpFactory = EnableDefaultFactory< ConcreteFactory, ConcreteLinOp, ParametersType, PolymorphicBase >
	This is an alias for the EnableDefaultFactory mixin, which correctly sets the template parameters to enable a subclass of LinOpFactory. More...

using	MachineTopology = machine_topology

template<typename T >
using	is_complex_s = detail::is_complex_impl< T >
	Allows to check if T is a complex value during compile time by accessing the `value` attribute of this struct. More...

template<typename T >
using	is_complex_or_scalar_s = detail::is_complex_or_scalar_impl< T >
	Allows to check if T is a complex or scalar value during compile time by accessing the `value` attribute of this struct. More...

template<typename T >
using	remove_complex = typename detail::remove_complex_s< T >::type
	Obtain the type which removed the complex of complex/scalar type or the template parameter of class by accessing the `type` attribute of this struct. More...

template<typename T >
using	to_complex = typename detail::to_complex_s< T >::type
	Obtain the type which adds the complex of complex/scalar type or the template parameter of class by accessing the `type` attribute of this struct. More...

template<typename T >
using	to_real = remove_complex< T >
	to_real is alias of remove_complex More...

template<typename T >
using	next_precision = typename detail::next_precision_impl< T >::type
	Obtains the next type in the singly-linked precision list.

template<typename T >
using	previous_precision = next_precision< T >
	Obtains the previous type in the singly-linked precision list. More...

template<typename T >
using	reduce_precision = typename detail::reduce_precision_impl< T >::type
	Obtains the next type in the hierarchy with lower precision than T.

template<typename T >
using	increase_precision = typename detail::increase_precision_impl< T >::type
	Obtains the next type in the hierarchy with higher precision than T.

template<typename... Ts>
using	highest_precision = typename detail::highest_precision_variadic< Ts... >::type
	Obtains the smallest arithmetic type that is able to store elements of all template parameter types exactly. More...

template<typename T , size_type Limit = sizeof(uint16) * byte_size>
using	truncate_type = std::conditional_t< detail::type_size_impl< T >::value >=2 *Limit, typename detail::truncate_type_impl< T >::type, T >
	Truncates the type by half (by dropping bits), but ensures that it is at least `Limit` bits wide.

using	size_type = std::size_t
	Integral type used for allocation quantities.

using	int8 = std::int8_t
	8-bit signed integral type.

using	int16 = std::int16_t
	16-bit signed integral type.

using	int32 = std::int32_t
	32-bit signed integral type.

using	int64 = std::int64_t
	64-bit signed integral type.

using	uint8 = std::uint8_t
	8-bit unsigned integral type.

using	uint16 = std::uint16_t
	16-bit unsigned integral type.

using	uint32 = std::uint32_t
	32-bit unsigned integral type.

using	uint64 = std::uint64_t
	64-bit unsigned integral type.

using	uintptr = std::uintptr_t
	Unsigned integer type capable of holding a pointer to void.

using	float16 = half
	Half precision floating point type.

using	float32 = float
	Single precision floating point type.

using	float64 = double
	Double precision floating point type.

using	full_precision = double
	The most precise floating-point type.

using	default_precision = double
	Precision used if no precision is explicitly specified.

Enumerations
enum	log_propagation_mode { log_propagation_mode::never, log_propagation_mode::automatic }
	How Logger events are propagated to their Executor. More...

enum	allocation_mode { device, unified_global, unified_host }
	Specify the mode of allocation for CUDA/HIP GPUs. More...

enum	layout_type { layout_type::array, layout_type::coordinate }
	Specifies the layout type when writing data in matrix market format. More...

Functions
template<typename ValueType >
ValueType	reduce_add (const array< ValueType > &input_arr, const ValueType init_val=0)
	Reduce (sum) the values in the array. More...

template<typename ValueType >
void	reduce_add (const array< ValueType > &input_arr, array< ValueType > &result)
	Reduce (sum) the values in the array. More...

template<typename ValueType >
array< ValueType >	make_array_view (std::shared_ptr< const Executor > exec, size_type size, ValueType *data)
	Helper function to create an array view deducing the value type. More...

template<typename ValueType >
detail::const_array_view< ValueType >	make_const_array_view (std::shared_ptr< const Executor > exec, size_type size, const ValueType *data)
	Helper function to create a const array view deducing the value type. More...

template<typename DimensionType >
batch_dim< 2, DimensionType >	transpose (const batch_dim< 2, DimensionType > &input)
	Returns a batch_dim object with its dimensions swapped for batched operators. More...

template<typename DimensionType >
constexpr dim< 2, DimensionType >	transpose (const dim< 2, DimensionType > &dimensions) noexcept
	Returns a dim<2> object with its dimensions swapped. More...

template<typename T >
constexpr bool	is_complex ()
	Checks if T is a complex type. More...

template<typename T >
constexpr bool	is_complex_or_scalar ()
	Checks if T is a complex/scalar type. More...

template<typename T >
constexpr reduce_precision< T >	round_down (T val)
	Reduces the precision of the input parameter. More...

template<typename T >
constexpr increase_precision< T >	round_up (T val)
	Increases the precision of the input parameter. More...

constexpr int64	ceildiv (int64 num, int64 den)
	Performs integer division with rounding up. More...

template<typename T >
constexpr T	zero ()
	Returns the additive identity for T. More...

template<typename T >
constexpr T	zero (const T &)
	Returns the additive identity for T. More...

template<typename T >
constexpr T	one ()
	Returns the multiplicative identity for T. More...

template<typename T >
constexpr T	one (const T &)
	Returns the multiplicative identity for T. More...

template<typename T >
constexpr bool	is_zero (T value)
	Returns true if and only if the given value is zero. More...

template<typename T >
constexpr bool	is_nonzero (T value)
	Returns true if and only if the given value is not zero. More...

template<typename T >
constexpr T	max (const T &x, const T &y)
	Returns the larger of the arguments. More...

template<typename T >
constexpr T	min (const T &x, const T &y)
	Returns the smaller of the arguments. More...

template<typename T >
constexpr auto	real (const T &x)
	Returns the real part of the object. More...

template<typename T >
constexpr auto	imag (const T &x)
	Returns the imaginary part of the object. More...

template<typename T >
constexpr auto	conj (const T &x)
	Returns the conjugate of an object. More...

template<typename T >
constexpr auto	squared_norm (const T &x) -> decltype(real(conj(x) *x))
	Returns the squared norm of the object. More...

template<typename T >
constexpr xstd::enable_if_t<!is_complex_s< T >::value, T >	abs (const T &x)
	Returns the absolute value of the object. More...

template<typename T >
constexpr xstd::enable_if_t< is_complex_s< T >::value, remove_complex< T > >	abs (const T &x)

template<typename T >
constexpr T	pi ()
	Returns the value of pi. More...

template<typename T >
constexpr std::complex< remove_complex< T > >	unit_root (int64 n, int64 k=1)
	Returns the value of exp(2 * pi * i * k / n), i.e. More...

template<typename T >
constexpr uint32	get_significant_bit (const T &n, uint32 hint=0u) noexcept
	Returns the position of the most significant bit of the number. More...

template<typename T >
constexpr T	get_superior_power (const T &base, const T &limit, const T &hint=T{1}) noexcept
	Returns the smallest power of `base` not smaller than `limit`. More...

template<typename T >
std::enable_if_t<!is_complex_s< T >::value, bool >	is_finite (const T &value)
	Checks if a floating point number is finite, meaning it is neither +/- infinity nor NaN. More...

template<typename T >
std::enable_if_t< is_complex_s< T >::value, bool >	is_finite (const T &value)
	Checks if all components of a complex value are finite, meaning they are neither +/- infinity nor NaN. More...

template<typename T >
T	safe_divide (T a, T b)
	Computes the quotient of the given parameters, guarding against division by zero. More...

template<typename T >
std::enable_if_t<!is_complex_s< T >::value, bool >	is_nan (const T &value)
	Checks if a floating point number is NaN. More...

template<typename T >
std::enable_if_t< is_complex_s< T >::value, bool >	is_nan (const T &value)
	Checks if any component of a complex value is NaN. More...

template<typename T >
constexpr std::enable_if_t<!is_complex_s< T >::value, T >	nan ()
	Returns a quiet NaN of the given type. More...

template<typename T >
constexpr std::enable_if_t< is_complex_s< T >::value, T >	nan ()
	Returns a complex with both components quiet NaN. More...

template<typename ValueType = default_precision, typename IndexType = int32>
matrix_data< ValueType, IndexType >	read_raw (std::istream &is)
	Reads a matrix stored in matrix market format from an input stream. More...

template<typename ValueType = default_precision, typename IndexType = int32>
matrix_data< ValueType, IndexType >	read_binary_raw (std::istream &is)
	Reads a matrix stored in Ginkgo's binary matrix format from an input stream. More...

template<typename ValueType = default_precision, typename IndexType = int32>
matrix_data< ValueType, IndexType >	read_generic_raw (std::istream &is)
	Reads a matrix stored in either binary or matrix market format from an input stream. More...

template<typename ValueType , typename IndexType >
void	write_raw (std::ostream &os, const matrix_data< ValueType, IndexType > &data, layout_type layout=layout_type::coordinate)
	Writes a matrix_data structure to a stream in matrix market format. More...

template<typename ValueType , typename IndexType >
void	write_binary_raw (std::ostream &os, const matrix_data< ValueType, IndexType > &data)
	Writes a matrix_data structure to a stream in binary format. More...

template<typename MatrixType , typename StreamType , typename... MatrixArgs>
std::unique_ptr< MatrixType >	read (StreamType &&is, MatrixArgs &&... args)
	Reads a matrix stored in matrix market format from an input stream. More...

template<typename MatrixType , typename StreamType , typename... MatrixArgs>
std::unique_ptr< MatrixType >	read_binary (StreamType &&is, MatrixArgs &&... args)
	Reads a matrix stored in binary format from an input stream. More...

template<typename MatrixType , typename StreamType , typename... MatrixArgs>
std::unique_ptr< MatrixType >	read_generic (StreamType &&is, MatrixArgs &&... args)
	Reads a matrix stored either in binary or matrix market format from an input stream. More...

template<typename MatrixPtrType , typename StreamType >
void	write (StreamType &&os, MatrixPtrType &&matrix, layout_type layout=detail::mtx_io_traits< std::remove_cv_t< detail::pointee< MatrixPtrType >>>::default_layout)
	Writes a matrix into an output stream in matrix market format. More...

template<typename MatrixPtrType , typename StreamType >
void	write_binary (StreamType &&os, MatrixPtrType &&matrix)
	Writes a matrix into an output stream in binary format. More...

template<typename R , typename T >
std::unique_ptr< R, std::function< void(R *)> >	copy_and_convert_to (std::shared_ptr< const Executor > exec, T *obj)
	Converts the object to R and places it on Executor exec. More...

template<typename R , typename T >
std::unique_ptr< const R, std::function< void(const R *)> >	copy_and_convert_to (std::shared_ptr< const Executor > exec, const T *obj)
	Converts the object to R and places it on Executor exec. More...

template<typename R , typename T >
std::shared_ptr< R >	copy_and_convert_to (std::shared_ptr< const Executor > exec, std::shared_ptr< T > obj)
	Converts the object to R and places it on Executor exec. More...

template<typename R , typename T >
std::shared_ptr< const R >	copy_and_convert_to (std::shared_ptr< const Executor > exec, std::shared_ptr< const T > obj)

template<typename ValueType , typename Ptr >
detail::temporary_conversion< std::conditional_t< std::is_const< detail::pointee< Ptr > >::value, const matrix::Dense< ValueType >, matrix::Dense< ValueType > > >	make_temporary_conversion (Ptr &&matrix)
	Convert the given LinOp from matrix::Dense<...> to matrix::Dense<ValueType>. More...

template<typename ValueType , typename Function , typename... Args>
void	precision_dispatch (Function fn, Args *... linops)
	Calls the given function with each given argument LinOp temporarily converted into matrix::Dense<ValueType> as parameters. More...

template<typename ValueType , typename Function >
void	precision_dispatch_real_complex (Function fn, const LinOp in, LinOp out)
	Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters. More...

template<typename ValueType , typename Function >
void	precision_dispatch_real_complex (Function fn, const LinOp alpha, const LinOp in, LinOp *out)
	Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters. More...

template<typename ValueType , typename Function >
void	precision_dispatch_real_complex (Function fn, const LinOp alpha, const LinOp in, const LinOp beta, LinOp out)
	Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters. More...

template<typename ValueType , typename Function >
void	mixed_precision_dispatch (Function fn, const LinOp in, LinOp out)
	Calls the given function with each given argument LinOp converted into matrix::Dense<ValueType> as parameters. More...

template<typename ValueType , typename Function , std::enable_if_t< is_complex< ValueType >()> * = nullptr>
void	mixed_precision_dispatch_real_complex (Function fn, const LinOp in, LinOp out)
	Calls the given function with the given LinOps cast to their dynamic type matrix::Dense<ValueType>* as parameters. More...

constexpr bool	operator< (const span &first, const span &second)

constexpr bool	operator<= (const span &first, const span &second)

constexpr bool	operator> (const span &first, const span &second)

constexpr bool	operator>= (const span &first, const span &second)

constexpr bool	operator== (const span &first, const span &second)

constexpr bool	operator!= (const span &first, const span &second)

template<typename Accessor >
constexpr range< accessor::unary_plus< Accessor > >	operator+ (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::unary_minus< Accessor > >	operator- (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::logical_not< Accessor > >	operator! (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::bitwise_not< Accessor > >	operator~ (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::zero_operation< Accessor > >	zero (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::one_operation< Accessor > >	one (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::abs_operation< Accessor > >	abs (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::real_operation< Accessor > >	real (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::imag_operation< Accessor > >	imag (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::conj_operation< Accessor > >	conj (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::squared_norm_operation< Accessor > >	squared_norm (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::transpose_operation< Accessor > >	transpose (const range< Accessor > &operand)

template<typename Accessor >
constexpr range< accessor::add< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator+ (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::add< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator+ (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::add< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator+ (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::add< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator+ (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::sub< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator- (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::sub< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator- (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::sub< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator- (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::sub< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator- (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::mul< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator* (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::mul< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator* (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::mul< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator* (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::mul< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator* (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::div< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator/ (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::div< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator/ (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::div< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator/ (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::div< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator/ (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::mod< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator% (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::mod< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator% (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::mod< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator% (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::mod< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator% (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::less< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator< (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::less< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator< (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::less< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator< (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::less< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator< (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::greater< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator> (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::greater< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator> (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::greater< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator> (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::greater< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator> (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::less_or_equal< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator<= (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::less_or_equal< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator<= (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::less_or_equal< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator<= (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::less_or_equal< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator<= (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::greater_or_equal< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator>= (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::greater_or_equal< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator>= (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::greater_or_equal< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator>= (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::greater_or_equal< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator>= (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::equal< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator== (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::equal< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator== (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::equal< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator== (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::equal< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator== (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::not_equal< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator!= (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::not_equal< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator!= (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::not_equal< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator!= (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::not_equal< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator!= (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::logical_or< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator\|\| (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::logical_or< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator\|\| (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::logical_or< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator\|\| (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::logical_or< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator\|\| (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::logical_and< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator&& (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::logical_and< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator&& (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::logical_and< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator&& (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::logical_and< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator&& (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::bitwise_or< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator\| (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::bitwise_or< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator\| (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::bitwise_or< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator\| (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::bitwise_or< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator\| (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::bitwise_and< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator& (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::bitwise_and< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator& (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::bitwise_and< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator& (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::bitwise_and< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator& (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::bitwise_xor< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator^ (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::bitwise_xor< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator^ (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::bitwise_xor< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator^ (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::bitwise_xor< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator^ (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::left_shift< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator<< (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::left_shift< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator<< (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::left_shift< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator<< (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::left_shift< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator<< (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::right_shift< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	operator>> (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::right_shift< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	operator>> (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::right_shift< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	operator>> (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::right_shift< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	operator>> (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::max_operation< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	max (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::max_operation< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	max (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::max_operation< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	max (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::max_operation< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	max (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::min_operation< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	min (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::min_operation< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	min (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::min_operation< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	min (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::min_operation< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	min (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Accessor >
constexpr range< accessor::mmul_operation< ::gko::detail::operation_kind::range_by_range, Accessor, Accessor > >	mmul (const range< Accessor > &first, const range< Accessor > &second)

template<typename FirstAccessor , typename SecondAccessor >
constexpr range< accessor::mmul_operation< ::gko::detail::operation_kind::range_by_range, FirstAccessor, SecondAccessor > >	mmul (const range< FirstAccessor > &first, const range< SecondAccessor > &second)

template<typename FirstAccessor , typename SecondOperand >
constexpr range< accessor::mmul_operation< ::gko::detail::operation_kind::range_by_scalar, FirstAccessor, SecondOperand > >	mmul (const range< FirstAccessor > &first, const SecondOperand &second)

template<typename FirstOperand , typename SecondAccessor >
constexpr range< accessor::mmul_operation< ::gko::detail::operation_kind::scalar_by_range, FirstOperand, SecondAccessor > >	mmul (const FirstOperand &first, const range< SecondAccessor > &second)

template<typename Ptr >
detail::temporary_clone< detail::pointee< Ptr > >	make_temporary_clone (std::shared_ptr< const Executor > exec, Ptr &&ptr)
	Creates a temporary_clone. More...

template<typename Ptr >
detail::temporary_clone< detail::pointee< Ptr > >	make_temporary_output_clone (std::shared_ptr< const Executor > exec, Ptr &&ptr)
	Creates a uninitialized temporary_clone that will be copied back to the input afterwards. More...

constexpr bool	operator== (precision_reduction x, precision_reduction y) noexcept
	Checks if two precision_reduction encodings are equal. More...

constexpr bool	operator!= (precision_reduction x, precision_reduction y) noexcept
	Checks if two precision_reduction encodings are different. More...

template<typename IndexType >
constexpr IndexType	invalid_index ()
	Value for an invalid signed index type.

template<typename Pointer >
detail::cloned_type< Pointer >	clone (const Pointer &p)
	Creates a unique clone of the object pointed to by `p`. More...

template<typename Pointer >
detail::cloned_type< Pointer >	clone (std::shared_ptr< const Executor > exec, const Pointer &p)
	Creates a unique clone of the object pointed to by `p` on Executor `exec`. More...

template<typename OwningPointer >
detail::shared_type< OwningPointer >	share (OwningPointer &&p)
	Marks the object pointed to by `p` as shared. More...

template<typename OwningPointer >
std::remove_reference< OwningPointer >::type &&	give (OwningPointer &&p)
	Marks that the object pointed to by `p` can be given to the callee. More...

template<typename Pointer >
std::enable_if< detail::have_ownership_s< Pointer >::value, detail::pointee< Pointer > * >::type	lend (const Pointer &p)
	Returns a non-owning (plain) pointer to the object pointed to by `p`. More...

template<typename Pointer >
std::enable_if<!detail::have_ownership_s< Pointer >::value, detail::pointee< Pointer > * >::type	lend (const Pointer &p)
	Returns a non-owning (plain) pointer to the object pointed to by `p`. More...

template<typename T , typename U >
std::decay_t< T > *	as (U *obj)
	Performs polymorphic type conversion. More...

template<typename T , typename U >
const std::decay_t< T > *	as (const U *obj)
	Performs polymorphic type conversion. More...

template<typename T , typename U >
std::decay_t< T > *	as (ptr_param< U > obj)
	Performs polymorphic type conversion on a ptr_param. More...

template<typename T , typename U >
const std::decay_t< T > *	as (ptr_param< const U > obj)
	Performs polymorphic type conversion. More...

template<typename T , typename U >
std::unique_ptr< std::decay_t< T > >	as (std::unique_ptr< U > &&obj)
	Performs polymorphic type conversion of a unique_ptr. More...

template<typename T , typename U >
std::shared_ptr< std::decay_t< T > >	as (std::shared_ptr< U > obj)
	Performs polymorphic type conversion of a shared_ptr. More...

template<typename T , typename U >
std::shared_ptr< const std::decay_t< T > >	as (std::shared_ptr< const U > obj)
	Performs polymorphic type conversion of a shared_ptr. More...

bool	operator== (const version &first, const version &second)

bool	operator!= (const version &first, const version &second)

bool	operator< (const version &first, const version &second)

bool	operator<= (const version &first, const version &second)

bool	operator> (const version &first, const version &second)

bool	operator>= (const version &first, const version &second)

std::ostream &	operator<< (std::ostream &os, const version &ver)
	Prints version information to a stream. More...

std::ostream &	operator<< (std::ostream &os, const version_info &ver_info)
	Prints library version information in human-readable format to a stream. More...

template<template< typename, typename > class MatrixType, typename... Args>
auto	with_matrix_type (Args &&... create_args)
	This function returns a type that delays a call to MatrixType::create. More...

template<typename VecPtr >
std::unique_ptr< matrix::Dense< typename detail::pointee< VecPtr >::value_type > >	make_dense_view (VecPtr &&vector)
	Creates a view of a given Dense vector. More...

template<typename VecPtr >
std::unique_ptr< const matrix::Dense< typename detail::pointee< VecPtr >::value_type > >	make_const_dense_view (VecPtr &&vector)
	Creates a view of a given Dense vector. More...

template<typename Matrix , typename... TArgs>
std::unique_ptr< Matrix >	initialize (size_type stride, std::initializer_list< typename Matrix::value_type > vals, std::shared_ptr< const Executor > exec, TArgs &&... create_args)
	Creates and initializes a column-vector. More...

template<typename Matrix , typename... TArgs>
std::unique_ptr< Matrix >	initialize (std::initializer_list< typename Matrix::value_type > vals, std::shared_ptr< const Executor > exec, TArgs &&... create_args)
	Creates and initializes a column-vector. More...

template<typename Matrix , typename... TArgs>
std::unique_ptr< Matrix >	initialize (size_type stride, std::initializer_list< std::initializer_list< typename Matrix::value_type >> vals, std::shared_ptr< const Executor > exec, TArgs &&... create_args)
	Creates and initializes a matrix. More...

template<typename Matrix , typename... TArgs>
std::unique_ptr< Matrix >	initialize (std::initializer_list< std::initializer_list< typename Matrix::value_type >> vals, std::shared_ptr< const Executor > exec, TArgs &&... create_args)
	Creates and initializes a matrix. More...

bool	operator== (const stopping_status &x, const stopping_status &y) noexcept
	Checks if two stopping statuses are equivalent. More...

bool	operator!= (const stopping_status &x, const stopping_status &y) noexcept
	Checks if two stopping statuses are different. More...

Variables
constexpr allocation_mode	default_cuda_alloc_mode

constexpr allocation_mode	default_hip_alloc_mode

constexpr size_type	byte_size = CHAR_BIT
	Number of bits in a byte.

Detailed Description

The Ginkgo namespace.

Typedef Documentation

◆ highest_precision

template<typename... Ts>

using gko::highest_precision = typedef typename detail::highest_precision_variadic<Ts...>::type

Obtains the smallest arithmetic type that is able to store elements of all template parameter types exactly.

All template type parameters need to be either real or complex types, mixing them is not possible.

Formally, it computes a right-fold over the type list, with the highest precision of a pair of real arithmetic types T1, T2 computed as decltype(T1{} + T2{}), or std::complex<highest_precision<remove_complex<T1>, remove_complex<T2>>> for complex types.

◆ is_complex_or_scalar_s

template<typename T >

using gko::is_complex_or_scalar_s = typedef detail::is_complex_or_scalar_impl<T>

Allows to check if T is a complex or scalar value during compile time by accessing the value attribute of this struct.

If value is true, T is a complex/scalar type, if it is false, T is not a complex/scalar type.

Template Parameters

T	type to check

◆ is_complex_s

template<typename T >

using gko::is_complex_s = typedef detail::is_complex_impl<T>

Allows to check if T is a complex value during compile time by accessing the value attribute of this struct.

If value is true, T is a complex type, if it is false, T is not a complex type.

Template Parameters

T	type to check

◆ previous_precision

template<typename T >

using gko::previous_precision = typedef next_precision<T>

Obtains the previous type in the singly-linked precision list.

Note: Currently our lists contains only two elements, so this is the same as next_precision.

◆ remove_complex

template<typename T >

using gko::remove_complex = typedef typename detail::remove_complex_s<T>::type

Obtain the type which removed the complex of complex/scalar type or the template parameter of class by accessing the type attribute of this struct.

Template Parameters

T	type to remove complex

Note: remove_complex<class> can not be used in friend class declaration.

◆ to_complex

template<typename T >

using gko::to_complex = typedef typename detail::to_complex_s<T>::type

Obtain the type which adds the complex of complex/scalar type or the template parameter of class by accessing the type attribute of this struct.

Template Parameters

T	type to complex_type

Note: to_complex<class> can not be used in friend class declaration. the followings are the error message from different combination. friend to_complex<Csr>; error: can not recognize it is class correctly. friend class to_complex<Csr>; error: using alias template specialization friend class to_complex_s<Csr<ValueType,IndexType>>::type; error: can not recognize it is class correctly.

◆ to_real

template<typename T >

using gko::to_real = typedef remove_complex<T>

to_real is alias of remove_complex

Template Parameters

T	type to real

Enumeration Type Documentation

◆ allocation_mode

enum gko::allocation_mode

strong

Specify the mode of allocation for CUDA/HIP GPUs.

device allocates memory on the device and Unified Memory model is not used.

unified_global allocates memory on the device, but is accessible by the host through the Unified memory model.

unified_host allocates memory on the host and it is not available on devices which do not have concurrent accesses switched on, but this access can be explicitly switched on, when necessary.

◆ layout_type

enum gko::layout_type

strong

Specifies the layout type when writing data in matrix market format.

Enumerator
array	The matrix should be written as dense matrix in column-major order.
coordinate	The matrix should be written as a sparse matrix in coordinate format.

◆ log_propagation_mode

enum gko::log_propagation_mode

strong

How Logger events are propagated to their Executor.

Enumerator

never

Events only get reported at loggers attached to the triggering object.

(Except for allocation/free, copy and Operations, since they happen at the Executor).

automatic

Events get reported to loggers attached to the triggering object and propagating loggers (Logger::needs_propagation() return true) attached to its Executor.

Function Documentation

◆ abs()

template<typename T >

constexpr xstd::enable_if_t<!is_complex_s<T>::value, T> gko::abs ( const T & x )

inlineconstexpr

Returns the absolute value of the object.

Template Parameters

T	the type of the object

Parameters

x	the object

Returns: x >= zero<T>() ? x : -x;

Referenced by is_finite().

◆ as() [1/7]

template<typename T , typename U >

const std::decay_t<T>* gko::as ( const U * obj )

inline

Performs polymorphic type conversion.

This is the constant version of the function.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the object which should be converted

Returns: If successful, returns a pointer to the subtype, otherwise throws NotSupported.

◆ as() [2/7]

template<typename T , typename U >

const std::decay_t<T>* gko::as ( ptr_param< const U > obj )

inline

Performs polymorphic type conversion.

This is the constant version of the function.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the object which should be converted

Returns: If successful, returns a pointer to the subtype, otherwise throws NotSupported.

References gko::ptr_param< T >::get().

◆ as() [3/7]

template<typename T , typename U >

std::decay_t<T>* gko::as ( ptr_param< U > obj )

inline

Performs polymorphic type conversion on a ptr_param.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the object which should be converted

Returns: If successful, returns a pointer to the subtype, otherwise throws NotSupported.

References gko::ptr_param< T >::get().

◆ as() [4/7]

template<typename T , typename U >

std::shared_ptr<const std::decay_t<T> > gko::as ( std::shared_ptr< const U > obj )

inline

Performs polymorphic type conversion of a shared_ptr.

This is the constant version of the function.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the shared_ptr to the object which should be converted.

Returns: If successful, returns a shared_ptr to the subtype, otherwise throws NotSupported. This pointer shares ownership with the input pointer.

◆ as() [5/7]

template<typename T , typename U >

std::shared_ptr<std::decay_t<T> > gko::as ( std::shared_ptr< U > obj )

inline

Performs polymorphic type conversion of a shared_ptr.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the shared_ptr to the object which should be converted.

Returns: If successful, returns a shared_ptr to the subtype, otherwise throws NotSupported. This pointer shares ownership with the input pointer.

◆ as() [6/7]

template<typename T , typename U >

std::unique_ptr<std::decay_t<T> > gko::as ( std::unique_ptr< U > && obj )

inline

Performs polymorphic type conversion of a unique_ptr.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the unique_ptr to the object which should be converted. If successful, it will be reset to a nullptr.

Returns: If successful, returns a unique_ptr to the subtype, otherwise throws NotSupported.

◆ as() [7/7]

template<typename T , typename U >

std::decay_t<T>* gko::as ( U * obj )

inline

Performs polymorphic type conversion.

Template Parameters

T	requested result type
U	static type of the passed object

Parameters

obj	the object which should be converted

Returns: If successful, returns a pointer to the subtype, otherwise throws NotSupported.

Referenced by gko::preconditioner::Isai< IsaiType, ValueType, IndexType >::get_approximate_inverse().

◆ ceildiv()

constexpr int64 gko::ceildiv	(	int64	num,
		int64	den
	)

inlineconstexpr

Performs integer division with rounding up.

Parameters

num	numerator
den	denominator

Returns: returns the ceiled quotient.

Referenced by gko::matrix::Csr< ValueType, IndexType >::load_balance::clac_size(), gko::preconditioner::block_interleaved_storage_scheme< index_type >::compute_storage_space(), and gko::matrix::Csr< ValueType, IndexType >::load_balance::process().

◆ clone() [1/2]

template<typename Pointer >

detail::cloned_type<Pointer> gko::clone ( const Pointer & p )

inline

Creates a unique clone of the object pointed to by p.

The pointee (i.e. *p) needs to have a clone method that returns a std::unique_ptr in order for this method to work.

Template Parameters

Pointer type of pointer to the object (plain or smart pointer)

Parameters

p	a pointer to the object

Note: The difference between this function and directly calling LinOp::clone() is that this one preserves the static type of the object.

Referenced by gko::preconditioner::Ic< LSolverType, IndexType >::operator=(), gko::preconditioner::Ilu< LSolverType, USolverType, ReverseApply, IndexType >::operator=(), gko::solver::EnablePreconditionable< Bicg< ValueType > >::set_preconditioner(), and gko::solver::EnableIterativeBase< Bicg< ValueType > >::set_stop_criterion_factory().

◆ clone() [2/2]

template<typename Pointer >

detail::cloned_type<Pointer> gko::clone	(	std::shared_ptr< const Executor >	exec,
		const Pointer &	p
	)

inline

Creates a unique clone of the object pointed to by p on Executor exec.

The pointee (i.e. *p) needs to have a clone method that takes an executor and returns a std::unique_ptr in order for this method to work.

Template Parameters

Pointer type of pointer to the object (plain or smart pointer)

Parameters

exec	the executor where the cloned object should be stored
p	a pointer to the object

Note: The difference between this function and directly calling LinOp::clone() is that this one preserves the static type of the object.

◆ conj()

template<typename T >

constexpr auto gko::conj ( const T & x )

inlineconstexpr

Returns the conjugate of an object.

Parameters

x	the number to conjugate

Returns: conjugate of the object (by default, the object itself)

Referenced by squared_norm().

◆ copy_and_convert_to() [1/4]

template<typename R , typename T >

std::unique_ptr<const R, std::function<void(const R*)> > gko::copy_and_convert_to	(	std::shared_ptr< const Executor >	exec,
		const T *	obj
	)

Converts the object to R and places it on Executor exec.

If the object is already of the requested type and on the requested executor, the copy and conversion is avoided and a reference to the original object is returned instead.

Template Parameters

R	the type to which the object should be converted
T	the type of the input object

Parameters

exec	the executor where the result should be placed
obj	the object that should be converted

Returns: a unique pointer (with dynamically bound deleter) to the converted object

Note: This is a version of the function which adds the const qualifier to the result if the input had the same qualifier.

◆ copy_and_convert_to() [2/4]

template<typename R , typename T >

std::shared_ptr<const R> gko::copy_and_convert_to	(	std::shared_ptr< const Executor >	exec,
		std::shared_ptr< const T >	obj
	)

This is the version that takes in the std::shared_ptr and returns a std::shared_ptr

If the object is already of the requested type and on the requested executor, the copy and conversion is avoided and a reference to the original object is returned instead.

Template Parameters

R	the type to which the object should be converted
T	the type of the input object

Parameters

exec	the executor where the result should be placed
obj	the object that should be converted

Returns: a shared pointer to the converted object

Note: This is a version of the function which adds the const qualifier to the result if the input had the same qualifier.

◆ copy_and_convert_to() [3/4]

template<typename R , typename T >

std::shared_ptr<R> gko::copy_and_convert_to	(	std::shared_ptr< const Executor >	exec,
		std::shared_ptr< T >	obj
	)

Converts the object to R and places it on Executor exec.

This is the version that takes in the std::shared_ptr and returns a std::shared_ptr

If the object is already of the requested type and on the requested executor, the copy and conversion is avoided and a reference to the original object is returned instead.

Template Parameters

R	the type to which the object should be converted
T	the type of the input object

Parameters

exec	the executor where the result should be placed
obj	the object that should be converted

Returns: a shared pointer to the converted object

◆ copy_and_convert_to() [4/4]

template<typename R , typename T >

std::unique_ptr<R, std::function<void(R*)> > gko::copy_and_convert_to	(	std::shared_ptr< const Executor >	exec,
		T *	obj
	)

Converts the object to R and places it on Executor exec.

If the object is already of the requested type and on the requested executor, the copy and conversion is avoided and a reference to the original object is returned instead.

Template Parameters

R	the type to which the object should be converted
T	the type of the input object

Parameters

exec	the executor where the result should be placed
obj	the object that should be converted

Returns: a unique pointer (with dynamically bound deleter) to the converted object

◆ get_significant_bit()

template<typename T >

constexpr uint32 gko::get_significant_bit	(	const T &	n,
		uint32	hint = `0u`
	)

constexprnoexcept

Returns the position of the most significant bit of the number.

This is the same as the rounded down base-2 logarithm of the number.

Template Parameters

T	a numeric type supporting bit shift and comparison

Parameters

n	a number
hint	a lower bound for the position o the significant bit

Returns: maximum of hint and the significant bit position of n

◆ get_superior_power()

template<typename T >

constexpr T gko::get_superior_power	(	const T &	base,
		const T &	limit,
		const T &	hint = `T{1}`
	)

constexprnoexcept

Returns the smallest power of base not smaller than limit.

Template Parameters

T	a numeric type supporting multiplication and comparison

Parameters

base	the base of the power to be returned
limit	the lower limit on the size of the power returned
hint	a lower bound on the result, has to be a power of base

Returns: the smallest power of base not smaller than limit

◆ give()

template<typename OwningPointer >

std::remove_reference<OwningPointer>::type&& gko::give ( OwningPointer && p )

inline

Marks that the object pointed to by p can be given to the callee.

Effectively calls std::move(p).

Template Parameters

OwningPointer type of pointer with ownership to the object (has to be a smart pointer)

Parameters

p	a pointer to the object

Note: The original pointer p becomes invalid after this call.

◆ imag()

template<typename T >

constexpr auto gko::imag ( const T & x )

inlineconstexpr

Returns the imaginary part of the object.

Template Parameters

T	type of the object

Parameters

x	the object

Returns: imaginary part of the object (by default, zero<T>())

◆ is_complex()

template<typename T >

constexpr bool gko::is_complex ( )

inlineconstexpr

Checks if T is a complex type.

Template Parameters

T	type to check

Returns: true if T is a complex type, false otherwise

◆ is_complex_or_scalar()

template<typename T >

constexpr bool gko::is_complex_or_scalar ( )

inlineconstexpr

Checks if T is a complex/scalar type.

Template Parameters

T	type to check

Returns: true if T is a complex/scalar type, false otherwise

◆ is_finite() [1/2]

template<typename T >

std::enable_if_t<!is_complex_s<T>::value, bool> gko::is_finite ( const T & value )

inline

Checks if a floating point number is finite, meaning it is neither +/- infinity nor NaN.

Template Parameters

T	type of the value to check

Parameters

value value to check

Returns: true if the value is finite, meaning it are neither +/- infinity nor NaN.

References abs().

Referenced by is_finite().

◆ is_finite() [2/2]

template<typename T >

std::enable_if_t<is_complex_s<T>::value, bool> gko::is_finite ( const T & value )

inline

Checks if all components of a complex value are finite, meaning they are neither +/- infinity nor NaN.

Template Parameters

T	complex type of the value to check

Parameters

value complex value to check

Returns: true if both components of the given value are finite, meaning they are neither +/- infinity nor NaN.

References is_finite().

◆ is_nan() [1/2]

template<typename T >

std::enable_if_t<!is_complex_s<T>::value, bool> gko::is_nan ( const T & value )

inline

Checks if a floating point number is NaN.

Template Parameters

T	type of the value to check

Parameters

value value to check

Returns: true if the value is NaN.

◆ is_nan() [2/2]

template<typename T >

std::enable_if_t<is_complex_s<T>::value, bool> gko::is_nan ( const T & value )

inline

Checks if any component of a complex value is NaN.

Template Parameters

T	complex type of the value to check

Parameters

value complex value to check

Returns: true if any component of the given value is NaN.

◆ is_nonzero()

template<typename T >

constexpr bool gko::is_nonzero ( T value )

inlineconstexpr

Returns true if and only if the given value is not zero.

Template Parameters

T	the type of the value

Parameters

value the given value

Returns: true iff the given value is not zero, i.e. value != zero<T>()

Referenced by gko::matrix_data< ValueType, IndexType >::diag().

◆ is_zero()

template<typename T >

constexpr bool gko::is_zero ( T value )

inlineconstexpr

Returns true if and only if the given value is zero.

Template Parameters

T	the type of the value

Parameters

value the given value

Returns: true iff the given value is zero, i.e. value == zero<T>()

Referenced by gko::matrix_data< ValueType, IndexType >::remove_zeros().

◆ lend() [1/2]

template<typename Pointer >

std::enable_if<detail::have_ownership_s<Pointer>::value, detail::pointee<Pointer>*>::type gko::lend ( const Pointer & p )

inline

Returns a non-owning (plain) pointer to the object pointed to by p.

Template Parameters

Pointer type of pointer to the object (plain or smart pointer)

Parameters

p	a pointer to the object

Note: This is the overload for owning (smart) pointers, that behaves the same as calling .get() on the smart pointer.

◆ lend() [2/2]

template<typename Pointer >

std::enable_if<!detail::have_ownership_s<Pointer>::value, detail::pointee<Pointer>*>::type gko::lend ( const Pointer & p )

inline

Returns a non-owning (plain) pointer to the object pointed to by p.

Template Parameters

Pointer type of pointer to the object (plain or smart pointer)

Parameters

p	a pointer to the object

Note: This is the overload for non-owning (plain) pointers, that just returns p.

◆ make_array_view()

template<typename ValueType >

array<ValueType> gko::make_array_view	(	std::shared_ptr< const Executor >	exec,
		size_type	size,
		ValueType *	data
	)

Helper function to create an array view deducing the value type.

Parameters

exec	the executor on which the array resides
size	the number of elements for the array
data	the pointer to the array we create a view on.

Template Parameters

ValueType the type of the array elements

Returns: array<ValueType>::view(exec, size, data)

References make_array_view().

Referenced by make_array_view().

◆ make_const_array_view()

template<typename ValueType >

detail::const_array_view<ValueType> gko::make_const_array_view	(	std::shared_ptr< const Executor >	exec,
		size_type	size,
		const ValueType *	data
	)

Helper function to create a const array view deducing the value type.

Parameters

exec	the executor on which the array resides
size	the number of elements for the array
data	the pointer to the array we create a view on.

Template Parameters

ValueType the type of the array elements

Returns: array<ValueType>::const_view(exec, size, data)

References make_const_array_view().

Referenced by make_const_array_view().

◆ make_const_dense_view()

template<typename VecPtr >

std::unique_ptr< const matrix::Dense<typename detail::pointee<VecPtr>::value_type> > gko::make_const_dense_view ( VecPtr && vector )

Creates a view of a given Dense vector.

Template Parameters

VecPtr a (smart or raw) pointer to the vector.

Parameters

vector the vector on which to create the view

References gko::matrix::Dense< ValueType >::create_const_view_of().

◆ make_dense_view()

template<typename VecPtr >

std::unique_ptr<matrix::Dense<typename detail::pointee<VecPtr>::value_type> > gko::make_dense_view ( VecPtr && vector )

Creates a view of a given Dense vector.

Template Parameters

VecPtr a (smart or raw) pointer to the vector.

Parameters

vector the vector on which to create the view

References gko::matrix::Dense< ValueType >::create_view_of().

◆ make_temporary_clone()

template<typename Ptr >

detail::temporary_clone<detail::pointee<Ptr> > gko::make_temporary_clone	(	std::shared_ptr< const Executor >	exec,
		Ptr &&	ptr
	)

Creates a temporary_clone.

This is a helper function which avoids the need to explicitly specify the type of the object, as would be the case if using the constructor of temporary_clone.

Parameters

exec	the executor where the clone will be created
ptr	a pointer to the object of which the clone will be created

Template Parameters

Ptr	the (raw or smart) pointer type to be temporarily cloned

Referenced by gko::ScaledIdentityAddable::add_scaled_identity(), gko::LinOp::apply(), gko::matrix::Csr< ValueType, IndexType >::inv_scale(), and gko::matrix::Csr< ValueType, IndexType >::scale().

◆ make_temporary_conversion()

template<typename ValueType , typename Ptr >

detail::temporary_conversion<std::conditional_t< std::is_const<detail::pointee<Ptr> >::value, const matrix::Dense<ValueType>, matrix::Dense<ValueType> > > gko::make_temporary_conversion ( Ptr && matrix )

Convert the given LinOp from matrix::Dense<...> to matrix::Dense<ValueType>.

The conversion tries to convert the input LinOp to all Dense types with value type recursively reachable by next_precision<...> starting from the ValueType template parameter. This means that all real-to-real and complex-to-complex conversions for default precisions are being considered. If the input matrix is non-const, the contents of the modified converted object will be converted back to the input matrix when the returned object is destroyed. This may lead to a loss of precision!

Parameters

matrix the input matrix which is supposed to be converted. It is wrapped unchanged if it is already of type matrix::Dense<ValueType>, otherwise it will be converted to this type if possible.

Returns: a detail::temporary_conversion pointing to the (potentially converted) object.

Exceptions

NotSupported if the input matrix cannot be converted to matrix::Dense<ValueType>

Template Parameters

ValueType the value type into whose associated matrix::Dense type to convert the input LinOp.

◆ make_temporary_output_clone()

template<typename Ptr >

detail::temporary_clone<detail::pointee<Ptr> > gko::make_temporary_output_clone	(	std::shared_ptr< const Executor >	exec,
		Ptr &&	ptr
	)

Creates a uninitialized temporary_clone that will be copied back to the input afterwards.

It can be used for output parameters to avoid an unnecessary copy in make_temporary_clone.

This is a helper function which avoids the need to explicitly specify the type of the object, as would be the case if using the constructor of temporary_clone.

Parameters

exec	the executor where the uninitialized clone will be created
ptr	a pointer to the object of which the clone will be created

Template Parameters

Ptr	the (raw or smart) pointer type to be temporarily cloned

◆ max()

template<typename T >

constexpr T gko::max	(	const T &	x,
		const T &	y
	)

inlineconstexpr

Returns the larger of the arguments.

Template Parameters

T	type of the arguments

Parameters

x	first argument
y	second argument

Returns: x >= y ? x : y

◆ min()

template<typename T >

constexpr T gko::min	(	const T &	x,
		const T &	y
	)

inlineconstexpr

Returns the smaller of the arguments.

Template Parameters

T	type of the arguments

Parameters

x	first argument
y	second argument

Returns: x <= y ? x : y

Referenced by gko::matrix::Csr< ValueType, IndexType >::load_balance::clac_size().

◆ mixed_precision_dispatch()

template<typename ValueType , typename Function >

void gko::mixed_precision_dispatch	(	Function	fn,
		const LinOp *	in,
		LinOp *	out
	)

Calls the given function with each given argument LinOp converted into matrix::Dense<ValueType> as parameters.

If GINKGO_MIXED_PRECISION is defined, this means that the function will be called with its dynamic type as a static type, so the (templated/generic) function will be instantiated with all pairs of Dense<ValueType> and Dense<next_precision<ValueType>> parameter types, and the appropriate overload will be called based on the dynamic type of the parameter.

If GINKGO_MIXED_PRECISION is not defined, it will behave exactly like precision_dispatch.

Parameters

fn	the given function. It will be called with one const and one non-const matrix::Dense<...> parameter based on the dynamic type of the inputs (GINKGO_MIXED_PRECISION) or of type matrix::Dense<ValueType> (no GINKGO_MIXED_PRECISION).
in	The first parameter to be cast (GINKGO_MIXED_PRECISION) or converted (no GINKGO_MIXED_PRECISION) and used to call `fn`.
out	The second parameter to be cast (GINKGO_MIXED_PRECISION) or converted (no GINKGO_MIXED_PRECISION) and used to call `fn`.

Template Parameters

ValueType	the value type to use for the parameters of `fn` (no GINKGO_MIXED_PRECISION). With GINKGO_MIXED_PRECISION enabled, it only matters whether this type is complex or real.
Function	the function pointer, lambda or other functor type to call with the converted arguments.

◆ mixed_precision_dispatch_real_complex()

template<typename ValueType , typename Function , std::enable_if_t< is_complex< ValueType >()> * = nullptr>

void gko::mixed_precision_dispatch_real_complex	(	Function	fn,
		const LinOp *	in,
		LinOp *	out
	)

Calls the given function with the given LinOps cast to their dynamic type matrix::Dense<ValueType>* as parameters.

If ValueType is real and both in and out are complex, uses matrix::Dense::get_real_view() to convert them into real matrices after precision conversion.

See also: mixed_precision_dispatch()

◆ nan() [1/2]

template<typename T >

constexpr std::enable_if_t<!is_complex_s<T>::value, T> gko::nan ( )

inlineconstexpr

Returns a quiet NaN of the given type.

Template Parameters

T	the type of the object

Returns: NaN.

Referenced by nan().

◆ nan() [2/2]

template<typename T >

constexpr std::enable_if_t<is_complex_s<T>::value, T> gko::nan ( )

inlineconstexpr

Returns a complex with both components quiet NaN.

Template Parameters

T	the type of the object

Returns: complex{NaN, NaN}.

References nan().

◆ one() [1/2]

template<typename T >

constexpr T gko::one ( )

inlineconstexpr

Returns the multiplicative identity for T.

Returns: the multiplicative identity for T

Referenced by unit_root().

◆ one() [2/2]

template<typename T >

constexpr T gko::one ( const T & )

inlineconstexpr

Returns the multiplicative identity for T.

Returns: the multiplicative identity for T

Note: This version takes an unused reference argument to avoid complicated calls like one<decltype(x)>(). Instead, it allows one(x).

◆ operator!=() [1/2]

bool gko::operator!=	(	const stopping_status &	x,
		const stopping_status &	y
	)

inlinenoexcept

Checks if two stopping statuses are different.

Parameters

x	a stopping status
y	a stopping status

Returns: true if and only if !(x == y)

◆ operator!=() [2/2]

constexpr bool gko::operator!=	(	precision_reduction	x,
		precision_reduction	y
	)

constexprnoexcept

Checks if two precision_reduction encodings are different.

Parameters

x	an encoding
y	an encoding

Returns: true if and only if x and y are different encodings.

◆ operator<<() [1/2]

std::ostream& gko::operator<<	(	std::ostream &	os,
		const version &	ver
	)

inline

Prints version information to a stream.

Parameters

os	output stream
ver	version structure

Returns: os

References gko::version::major, gko::version::minor, gko::version::patch, and gko::version::tag.

◆ operator<<() [2/2]

std::ostream& gko::operator<<	(	std::ostream &	os,
		const version_info &	ver_info
	)

Prints library version information in human-readable format to a stream.

Parameters

os	output stream
ver_info	version information

Returns: os

◆ operator==() [1/2]

bool gko::operator==	(	const stopping_status &	x,
		const stopping_status &	y
	)

inlinenoexcept

Checks if two stopping statuses are equivalent.

Parameters

x	a stopping status
y	a stopping status

Returns: true if and only if both x and y have the same mask and converged and finalized state

◆ operator==() [2/2]

constexpr bool gko::operator==	(	precision_reduction	x,
		precision_reduction	y
	)

constexprnoexcept

Checks if two precision_reduction encodings are equal.

Parameters

x	an encoding
y	an encoding

Returns: true if and only if x and y are the same encodings

◆ pi()

template<typename T >

constexpr T gko::pi ( )

inlineconstexpr

Returns the value of pi.

Template Parameters

T	the value type to return

◆ precision_dispatch()

template<typename ValueType , typename Function , typename... Args>

void gko::precision_dispatch	(	Function	fn,
		Args *...	linops
	)

Calls the given function with each given argument LinOp temporarily converted into matrix::Dense<ValueType> as parameters.

Parameters

fn	the given function. It will be passed one (potentially const) matrix::Dense<ValueType>* parameter per parameter in the parameter pack `linops`.
linops	the given arguments to be converted and passed on to fn.

Template Parameters

ValueType	the value type to use for the parameters of `fn`.
Function	the function pointer, lambda or other functor type to call with the converted arguments.
Args	the argument type list.

◆ precision_dispatch_real_complex() [1/3]

template<typename ValueType , typename Function >

void gko::precision_dispatch_real_complex	(	Function	fn,
		const LinOp *	alpha,
		const LinOp *	in,
		const LinOp *	beta,
		LinOp *	out
	)

Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters.

If ValueType is real and both in and out are complex, uses matrix::Dense::get_real_view() to convert them into real matrices after precision conversion.

See also: precision_dispatch()

◆ precision_dispatch_real_complex() [2/3]

template<typename ValueType , typename Function >

void gko::precision_dispatch_real_complex	(	Function	fn,
		const LinOp *	alpha,
		const LinOp *	in,
		LinOp *	out
	)

Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters.

If ValueType is real and both in and out are complex, uses matrix::Dense::get_real_view() to convert them into real matrices after precision conversion.

See also: precision_dispatch()

◆ precision_dispatch_real_complex() [3/3]

template<typename ValueType , typename Function >

void gko::precision_dispatch_real_complex	(	Function	fn,
		const LinOp *	in,
		LinOp *	out
	)

Calls the given function with the given LinOps temporarily converted to matrix::Dense<ValueType>* as parameters.

If ValueType is real and both input vectors are complex, uses matrix::Dense::get_real_view() to convert them into real matrices after precision conversion.

See also: precision_dispatch()

◆ read()

template<typename MatrixType , typename StreamType , typename... MatrixArgs>

std::unique_ptr<MatrixType> gko::read	(	StreamType &&	is,
		MatrixArgs &&...	args
	)

inline

Reads a matrix stored in matrix market format from an input stream.

Template Parameters

MatrixType	a ReadableFromMatrixData LinOp type used to store the matrix once it's been read from disk.
StreamType	type of stream used to write the data to
MatrixArgs	additional argument types passed to MatrixType constructor

Parameters

is	input stream from which to read the data
args	additional arguments passed to MatrixType constructor

Returns: A MatrixType LinOp filled with data from filename

References read_raw().

Referenced by gko::ReadableFromMatrixData< ValueType, int32 >::read().

◆ read_binary()

template<typename MatrixType , typename StreamType , typename... MatrixArgs>

std::unique_ptr<MatrixType> gko::read_binary	(	StreamType &&	is,
		MatrixArgs &&...	args
	)

inline

Reads a matrix stored in binary format from an input stream.

Template Parameters

MatrixType	a ReadableFromMatrixData LinOp type used to store the matrix once it's been read from disk.
StreamType	type of stream used to write the data to
MatrixArgs	additional argument types passed to MatrixType constructor

Parameters

is	input stream from which to read the data
args	additional arguments passed to MatrixType constructor

Returns: A MatrixType LinOp filled with data from filename

References read_binary_raw().

◆ read_binary_raw()

template<typename ValueType = default_precision, typename IndexType = int32>

matrix_data<ValueType, IndexType> gko::read_binary_raw ( std::istream & is )

Reads a matrix stored in Ginkgo's binary matrix format from an input stream.

Note that this format depends on the processor's endianness, so files from a big endian processor can't be read from a little endian processor and vice-versa.

The binary format has the following structure (in system endianness):

A 32 byte header consisting of 4 uint64_t values: magic = GINKGO__: The highest two bytes stand for value and index type. value type: S (float), D (double), C (complex<float>), Z(complex<double>) index type: I (int32), L (int64) num_rows: Number of rows num_cols: Number of columns num_entries: Number of (row, column, value) tuples to follow
Following are num_entries blocks of size sizeof(IndexType) * 2 + sizeof(ValueType). Each consists of a row index stored as IndexType, followed by a column index stored as IndexType and a value stored as ValueType.

Template Parameters

ValueType	type of matrix values
IndexType	type of matrix indexes

Parameters

is	input stream from which to read the data

Returns: A matrix_data structure containing the matrix. The nonzero elements are sorted in lexicographic order of their (row, column) indexes.

Note: This is an advanced routine that will return the raw matrix data structure. Consider using gko::read_binary instead.

Referenced by read_binary().

◆ read_generic()

template<typename MatrixType , typename StreamType , typename... MatrixArgs>

std::unique_ptr<MatrixType> gko::read_generic	(	StreamType &&	is,
		MatrixArgs &&...	args
	)

inline

Reads a matrix stored either in binary or matrix market format from an input stream.

Template Parameters

MatrixType	a ReadableFromMatrixData LinOp type used to store the matrix once it's been read from disk.
StreamType	type of stream used to write the data to
MatrixArgs	additional argument types passed to MatrixType constructor

Parameters

is	input stream from which to read the data
args	additional arguments passed to MatrixType constructor

Returns: A MatrixType LinOp filled with data from filename

References read_generic_raw().

◆ read_generic_raw()

template<typename ValueType = default_precision, typename IndexType = int32>

matrix_data<ValueType, IndexType> gko::read_generic_raw ( std::istream & is )

Reads a matrix stored in either binary or matrix market format from an input stream.

Template Parameters

ValueType	type of matrix values
IndexType	type of matrix indexes

Parameters

is	input stream from which to read the data

Returns: A matrix_data structure containing the matrix. The nonzero elements are sorted in lexicographic order of their (row, column) indexes.

Note: This is an advanced routine that will return the raw matrix data structure. Consider using gko::read_generic instead.

Referenced by read_generic().

◆ read_raw()

template<typename ValueType = default_precision, typename IndexType = int32>

matrix_data<ValueType, IndexType> gko::read_raw ( std::istream & is )

Reads a matrix stored in matrix market format from an input stream.

Template Parameters

ValueType	type of matrix values
IndexType	type of matrix indexes

Parameters

is	input stream from which to read the data

Returns: A matrix_data structure containing the matrix. The nonzero elements are sorted in lexicographic order of their (row, column) indexes.

Note: This is an advanced routine that will return the raw matrix data structure. Consider using gko::read instead.

Referenced by read().

◆ real()

template<typename T >

constexpr auto gko::real ( const T & x )

inlineconstexpr

Returns the real part of the object.

Template Parameters

T	type of the object

Parameters

x	the object

Returns: real part of the object (by default, the object itself)

Referenced by squared_norm().

◆ reduce_add() [1/2]

template<typename ValueType >

void gko::reduce_add	(	const array< ValueType > &	input_arr,
		array< ValueType > &	result
	)

Reduce (sum) the values in the array.

Template Parameters

The	type of the input data

Parameters

[in]	input_arr	the input array to be reduced
[in,out]	result	the reduced value. The result is written into the first entry and the value in the first entry is used as the initial value for the reduce.

◆ reduce_add() [2/2]

template<typename ValueType >

ValueType gko::reduce_add	(	const array< ValueType > &	input_arr,
		const ValueType	init_val = `0`
	)

Reduce (sum) the values in the array.

Template Parameters

The	type of the input data

Parameters

[in]	input_arr	the input array to be reduced
[in]	init_val	the initial value

Returns: the reduced value

◆ round_down()

template<typename T >

constexpr reduce_precision<T> gko::round_down ( T val )

inlineconstexpr

Reduces the precision of the input parameter.

Template Parameters

T	the original precision

Parameters

val	the value to round down

Returns: the rounded down value

◆ round_up()

template<typename T >

constexpr increase_precision<T> gko::round_up ( T val )

inlineconstexpr

Increases the precision of the input parameter.

Template Parameters

T	the original precision

Parameters

val	the value to round up

Returns: the rounded up value

◆ safe_divide()

template<typename T >

T gko::safe_divide	(	T	a,
		T	b
	)

inline

Computes the quotient of the given parameters, guarding against division by zero.

Template Parameters

T	value type of the parameters

Parameters

a	the dividend
b	the divisor

Returns: the value of a / b if b is non-zero, zero otherwise.

◆ share()

template<typename OwningPointer >

detail::shared_type<OwningPointer> gko::share ( OwningPointer && p )

inline

Marks the object pointed to by p as shared.

Effectively converts a pointer with ownership to std::shared_ptr.

Template Parameters

OwningPointer type of pointer with ownership to the object (has to be a smart pointer)

Parameters

p	a pointer to the object. It must be a temporary or explicitly marked movable (rvalue reference).

Note: The original pointer p becomes invalid after this call.

Referenced by gko::preconditioner::Ic< LSolverType, IndexType >::conj_transpose(), gko::preconditioner::Ilu< LSolverType, USolverType, ReverseApply, IndexType >::conj_transpose(), gko::preconditioner::Ic< LSolverType, IndexType >::transpose(), and gko::preconditioner::Ilu< LSolverType, USolverType, ReverseApply, IndexType >::transpose().

◆ squared_norm()

template<typename T >

constexpr auto gko::squared_norm ( const T & x ) -> decltype(real(conj(x) * x))

inlineconstexpr

Returns the squared norm of the object.

Template Parameters

T	type of the object.

Returns: The squared norm of the object.

References conj(), and real().

◆ transpose() [1/2]

template<typename DimensionType >

batch_dim<2, DimensionType> gko::transpose ( const batch_dim< 2, DimensionType > & input )

inline

Returns a batch_dim object with its dimensions swapped for batched operators.

Template Parameters

DimensionType datatype used to represent each dimension

Parameters

dimensions original object

Returns: a batch_dim object with dimensions swapped

References gko::batch_dim< Dimensionality, DimensionType >::get_common_size(), and gko::batch_dim< Dimensionality, DimensionType >::get_num_batch_items().

Referenced by gko::preconditioner::Ic< LSolverType, IndexType >::conj_transpose(), gko::preconditioner::Ilu< LSolverType, USolverType, ReverseApply, IndexType >::conj_transpose(), gko::preconditioner::Ic< LSolverType, IndexType >::transpose(), and gko::preconditioner::Ilu< LSolverType, USolverType, ReverseApply, IndexType >::transpose().

◆ transpose() [2/2]

template<typename DimensionType >

constexpr dim<2, DimensionType> gko::transpose ( const dim< 2, DimensionType > & dimensions )

inlineconstexprnoexcept

Returns a dim<2> object with its dimensions swapped.

Template Parameters

DimensionType datatype used to represent each dimension

Parameters

dimensions original object

Returns: a dim<2> object with its dimensions swapped

◆ unit_root()

template<typename T >

constexpr std::complex<remove_complex<T> > gko::unit_root	(	int64	n,
		int64	k = `1`
	)

inlineconstexpr

Returns the value of exp(2 * pi * i * k / n), i.e.

an nth root of unity.

Parameters

n	the denominator of the argument
k	the numerator of the argument. Defaults to 1.

Template Parameters

T	the corresponding real value type.

References one().

◆ with_matrix_type()

template<template< typename, typename > class MatrixType, typename... Args>

auto gko::with_matrix_type ( Args &&... create_args )

This function returns a type that delays a call to MatrixType::create.

It can be used to set the used value and index type, as well as the executor at a later stage.

For example, the following code creates first a temporary object, which is then used later to construct an operator of the previously defined base type:

auto type = gko::with_matrix_type<gko::matrix::Csr>();
...
std::unique_ptr<LinOp> concrete_op
if(flag1){
  concrete_op = type.template create<double, int>(exec);
} else {
  concrete_op = type.template create<float, int>(exec);
}

Note: This is mainly a helper function to specify the local matrix type for a gko::experimental::distributed::Matrix more easily.

Template Parameters

MatrixType	A template type that accepts two types, the first one will be set to the value type, the second one to the index type.
Args	Types of the arguments passed to MatrixType::create.

Parameters

create_args arguments that will be forwarded to MatrixType::create

Returns: A type with a function create<value_type, index_type>(executor).

◆ write()

template<typename MatrixPtrType , typename StreamType >

void gko::write	(	StreamType &&	os,
		MatrixPtrType &&	matrix,
		layout_type	layout = `detail::mtx_io_traits< std::remove_cv_t<detail::pointee<MatrixPtrType>>>::default_layout`
	)

inline

Writes a matrix into an output stream in matrix market format.

Template Parameters

MatrixPtrType	a (smart or raw) pointer to a WritableToMatrixData object providing data to be written.
StreamType	type of stream used to write the data to

Parameters

os	output stream where the data is to be written
matrix	the matrix to write
layout	the layout used in the output

References write_raw().

◆ write_binary()

template<typename MatrixPtrType , typename StreamType >

void gko::write_binary	(	StreamType &&	os,
		MatrixPtrType &&	matrix
	)

inline

Writes a matrix into an output stream in binary format.

Note that this format depends on the processor's endianness, so files from a big endian processor can't be read from a little endian processor and vice-versa.

Template Parameters

MatrixPtrType	a (smart or raw) pointer to a WritableToMatrixData object providing data to be written.
StreamType	type of stream used to write the data to

Parameters

os	output stream where the data is to be written
matrix	the matrix to write

References write_binary_raw().

◆ write_binary_raw()

template<typename ValueType , typename IndexType >

void gko::write_binary_raw	(	std::ostream &	os,
		const matrix_data< ValueType, IndexType > &	data
	)

Writes a matrix_data structure to a stream in binary format.

Note that this format depends on the processor's endianness, so files from a big endian processor can't be read from a little endian processor and vice-versa.

Template Parameters

ValueType	type of matrix values
IndexType	type of matrix indexes

Parameters

os	output stream where the data is to be written
data	the matrix data to write

Note: This is an advanced routine that writes the raw matrix data structure. If you are trying to write an existing matrix, consider using gko::write_binary instead.

Referenced by write_binary().

◆ write_raw()

template<typename ValueType , typename IndexType >

void gko::write_raw	(	std::ostream &	os,
		const matrix_data< ValueType, IndexType > &	data,
		layout_type	layout = `layout_type::coordinate`
	)

Writes a matrix_data structure to a stream in matrix market format.

Template Parameters

ValueType	type of matrix values
IndexType	type of matrix indexes

Parameters

os	output stream where the data is to be written
data	the matrix data to write
layout	the layout used in the output

Note: This is an advanced routine that writes the raw matrix data structure. If you are trying to write an existing matrix, consider using gko::write instead.

Referenced by write().

◆ zero() [1/2]

template<typename T >

constexpr T gko::zero ( )

inlineconstexpr

Returns the additive identity for T.

Returns: additive identity for T

Referenced by gko::matrix::Hybrid< ValueType, IndexType >::strategy_type::strategy_type().

◆ zero() [2/2]

template<typename T >

constexpr T gko::zero ( const T & )

inlineconstexpr

Returns the additive identity for T.

Returns: additive identity for T

Note: This version takes an unused reference argument to avoid complicated calls like zero<decltype(x)>(). Instead, it allows zero(x).

Variable Documentation

◆ default_cuda_alloc_mode

constexpr allocation_mode gko::default_cuda_alloc_mode

constexpr

Initial value:

=

allocation_mode::unified_global

◆ default_hip_alloc_mode

constexpr allocation_mode gko::default_hip_alloc_mode

constexpr

Initial value:

=

allocation_mode::unified_global

Namespaces

Classes

Typedefs

Enumerations

Functions

Variables

Detailed Description

Typedef Documentation

◆ highest_precision

◆ is_complex_or_scalar_s

◆ is_complex_s

◆ previous_precision

◆ remove_complex

◆ to_complex

◆ to_real

Enumeration Type Documentation

◆ allocation_mode

◆ layout_type

◆ log_propagation_mode

Function Documentation

◆ abs()

◆ as() [1/7]

◆ as() [2/7]

◆ as() [3/7]

◆ as() [4/7]

◆ as() [5/7]

◆ as() [6/7]

◆ as() [7/7]

◆ ceildiv()

◆ clone() [1/2]

◆ clone() [2/2]

◆ conj()

◆ copy_and_convert_to() [1/4]

◆ copy_and_convert_to() [2/4]

◆ copy_and_convert_to() [3/4]

◆ copy_and_convert_to() [4/4]

◆ get_significant_bit()

◆ get_superior_power()

◆ give()

◆ imag()

◆ is_complex()

◆ is_complex_or_scalar()

◆ is_finite() [1/2]

◆ is_finite() [2/2]

◆ is_nan() [1/2]

◆ is_nan() [2/2]

◆ is_nonzero()

◆ is_zero()

◆ lend() [1/2]

◆ lend() [2/2]

◆ make_array_view()

◆ make_const_array_view()

◆ make_const_dense_view()

◆ make_dense_view()

◆ make_temporary_clone()

◆ make_temporary_conversion()

◆ make_temporary_output_clone()

◆ max()

◆ min()

◆ mixed_precision_dispatch()

◆ mixed_precision_dispatch_real_complex()

◆ nan() [1/2]

◆ nan() [2/2]

◆ one() [1/2]

◆ one() [2/2]

◆ operator!=() [1/2]

◆ operator!=() [2/2]

◆ operator<<() [1/2]

◆ operator<<() [2/2]

◆ operator==() [1/2]

◆ operator==() [2/2]

◆ pi()

◆ precision_dispatch()

◆ precision_dispatch_real_complex() [1/3]

◆ precision_dispatch_real_complex() [2/3]

◆ precision_dispatch_real_complex() [3/3]

◆ read()

◆ read_binary()

◆ read_binary_raw()

◆ read_generic()