The ginkgo Solve namespace. More...

Namespaces
	multigrid
	The solver multigrid namespace.

Classes
class	ApplyWithInitialGuess
	ApplyWithInitialGuess provides a way to give the input guess for apply function. More...

class	Bicg
	BICG or the Biconjugate gradient method is a Krylov subspace solver. More...

class	Bicgstab
	BiCGSTAB or the Bi-Conjugate Gradient-Stabilized is a Krylov subspace solver. More...

class	CbGmres
	CB-GMRES or the compressed basis generalized minimal residual method is an iterative type Krylov subspace method which is suitable for nonsymmetric linear systems. More...

class	Cg
	CG or the conjugate gradient method is an iterative type Krylov subspace method which is suitable for symmetric positive definite methods. More...

class	Cgs
	CGS or the conjugate gradient square method is an iterative type Krylov subspace method which is suitable for general systems. More...

struct	enable_iterative_solver_factory_parameters

struct	enable_preconditioned_iterative_solver_factory_parameters

class	EnableApplyWithInitialGuess
	EnableApplyWithInitialGuess providing default operation for ApplyWithInitialGuess with correct validation and log. More...

class	EnableIterativeBase
	A LinOp deriving from this CRTP class stores a stopping criterion factory and allows applying with a guess. More...

class	EnablePreconditionable
	Mixin providing default operation for Preconditionable with correct value semantics. More...

class	EnablePreconditionedIterativeSolver
	A LinOp implementing this interface stores a system matrix and stopping criterion factory. More...

class	EnableSolverBase
	A LinOp deriving from this CRTP class stores a system matrix. More...

class	Fcg
	FCG or the flexible conjugate gradient method is an iterative type Krylov subspace method which is suitable for symmetric positive definite methods. More...

class	Gcr
	GCR or the generalized conjugate residual method is an iterative type Krylov subspace method similar to GMRES which is suitable for nonsymmetric linear systems. More...

class	Gmres
	GMRES or the generalized minimal residual method is an iterative type Krylov subspace method which is suitable for nonsymmetric linear systems. More...

struct	has_with_criteria
	Helper structure to test if the Factory of SolverType has a function `with_criteria`. More...

struct	has_with_criteria< SolverType, std::void_t< decltype(SolverType::build().with_criteria(std::shared_ptr< const stop::CriterionFactory >()))> >
	Helper structure to test if the Factory of SolverType has a function `with_criteria`.

class	Idr
	IDR(s) is an efficient method for solving large nonsymmetric systems of linear equations. More...

class	Ir
	Iterative refinement (IR) is an iterative method that uses another coarse method to approximate the error of the current solution via the current residual. More...

class	IterativeBase
	A LinOp implementing this interface stores a stopping criterion factory. More...

class	LowerTrs
	LowerTrs is the triangular solver which solves the system L x = b, when L is a lower triangular matrix. More...

class	Multigrid
	Multigrid methods have a hierarchy of many levels, whose corase level is a subset of the fine level, of the problem. More...

class	SolverBase

class	UpperTrs
	UpperTrs is the triangular solver which solves the system U x = b, when U is an upper triangular matrix. More...

struct	workspace_traits
	Traits class providing information on the type and location of workspace vectors inside a solver. More...

struct	workspace_traits< Bicg< ValueType > >

struct	workspace_traits< Bicgstab< ValueType > >

struct	workspace_traits< Cg< ValueType > >

struct	workspace_traits< Cgs< ValueType > >

struct	workspace_traits< Fcg< ValueType > >

struct	workspace_traits< Gcr< ValueType > >

struct	workspace_traits< gko::experimental::solver::Direct< ValueType, IndexType > >

struct	workspace_traits< Gmres< ValueType > >

struct	workspace_traits< Idr< ValueType > >

struct	workspace_traits< Ir< ValueType > >

struct	workspace_traits< LowerTrs< ValueType, IndexType > >

struct	workspace_traits< Multigrid >

struct	workspace_traits< UpperTrs< ValueType, IndexType > >

Typedefs
template<typename ValueType = default_precision>
using	Richardson = Ir< ValueType >

Enumerations
enum	initial_guess_mode { initial_guess_mode::zero, initial_guess_mode::rhs, initial_guess_mode::provided }
	Give a initial guess mode about the input of the apply method. More...

enum	trisolve_algorithm { sparselib, syncfree }
	A helper for algorithm selection in the triangular solvers. More...

Functions
template<typename ValueType >
auto	build_smoother (std::shared_ptr< const LinOpFactory > factory, size_type iteration=1, ValueType relaxation_factor=0.9)
	build_smoother gives a shortcut to build a smoother by IR(Richardson) with limited stop criterion(iterations and relacation_factor). More...

template<typename ValueType >
auto	build_smoother (std::shared_ptr< const LinOp > solver, size_type iteration=1, ValueType relaxation_factor=0.9)
	build_smoother gives a shortcut to build a smoother by IR(Richardson) with limited stop criterion(iterations and relacation_factor). More...

Variables
constexpr size_type	gcr_default_krylov_dim = 100u

constexpr size_type	default_krylov_dim = 100u

constexpr size_type	gmres_default_krylov_dim = 100u

Detailed Description

The ginkgo Solve namespace.

The ginkgo Solver namespace.

Enumeration Type Documentation

◆ initial_guess_mode

enum gko::solver::initial_guess_mode

strong

Give a initial guess mode about the input of the apply method.

Enumerator
zero	the input is zero
rhs	the input is right hand side
provided	the input is provided

◆ trisolve_algorithm

enum gko::solver::trisolve_algorithm

strong

A helper for algorithm selection in the triangular solvers.

It currently only matters for the Cuda executor as there, we have a choice between the Ginkgo syncfree and cuSPARSE implementations.

Function Documentation

◆ build_smoother() [1/2]

template<typename ValueType >

auto gko::solver::build_smoother	(	std::shared_ptr< const LinOp >	solver,
		size_type	iteration = `1`,
		ValueType	relaxation_factor = `0.9`
	)

build_smoother gives a shortcut to build a smoother by IR(Richardson) with limited stop criterion(iterations and relacation_factor).

Parameters

solver	the shared pointer of solver
iteration	the maximum number of iteration, which default is 1
relaxation_factor	the relaxation factor for Richardson

Returns: the pointer of Ir(Richardson)

Note: this is the overload function for LinOp.

◆ build_smoother() [2/2]

template<typename ValueType >

auto gko::solver::build_smoother	(	std::shared_ptr< const LinOpFactory >	factory,
		size_type	iteration = `1`,
		ValueType	relaxation_factor = `0.9`
	)

build_smoother gives a shortcut to build a smoother by IR(Richardson) with limited stop criterion(iterations and relacation_factor).

Parameters

factory	the shared pointer of factory
iteration	the maximum number of iteration, which default is 1
relaxation_factor	the relaxation factor for Richardson

Returns: the pointer of Ir(Richardson)

Namespaces

Classes

Typedefs

Enumerations

Functions