AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ > Class Template Reference#

DiFfRG: DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ > Class Template Reference

DiFfRG

Discretization Framework for functional Renormalization Group flows

The abstract base class for all timestepping algorithms. It provides a standard constructor which populates typical timestepping parameters from a given JSONValue object, such as the timestep sizes, tolerances, verbosity, etc. that are used in the timestepping algorithms. More...

#include <abstract_timestepper.hh>

Classes
struct	ConsoleOutStats

struct	ExplicitParameters

struct	ImplicitParameters

Public Member Functions
	AbstractTimestepper (const JSONValue &json, AbstractAssembler< VectorType, SparseMatrixType, dim > assembler, DataOutput< dim, VectorType > data_out=nullptr, AbstractAdaptor< VectorType > *adaptor=nullptr)
	Construct a new Abstract Timestepper object.

DataOutput< dim, VectorType > *	get_data_out ()
	Utility function to obtain a DataOutput object. If no DataOutput object is provided, a default one is created.

AbstractAdaptor< VectorType > *	get_adaptor ()
	Utility function to obtain an Adaptor object. If no Adaptor object is provided, a default one is created, which is the NoAdaptivity object, i.e. no mesh adaptivity is used.

virtual void	run (AbstractFlowingVariables< NumberType > *initial_condition, const double t_start, const double t_stop)=0
	Any derived class must implement this method to run the timestepping algorithm.

Protected Types
using	VectorType = VectorType_

using	NumberType = typename get_type::NumberType< VectorType >

using	SparseMatrixType = SparseMatrixType_

using	InverseSparseMatrixType = typename get_type::InverseSparseMatrixType< SparseMatrixType >

using	BlockVectorType = dealii::BlockVector< NumberType >

Protected Member Functions
void	print_console_out (const std::string &name, const double t, const size_t milliseconds, const std::optional< std::string > calc_dt=std::nullopt, const size_t calls=0) const

void	console_out (const double t, const std::string name, const int verbosity_level, const double calc_dt_ms=-1.0) const
	Pretty-print the status of the timestepping algorithm to the console.

Static Protected Member Functions
static std::string	console_out_category (const std::string &name)

static std::string	format_uncertain_ms (const double mean_ms, const double uncertainty_ms)

Protected Attributes
const JSONValue	json

AbstractAssembler< VectorType, SparseMatrixType, dim > *	assembler

DataOutput< dim, VectorType > *	data_out

AbstractAdaptor< VectorType > *	adaptor

const std::chrono::time_point< std::chrono::high_resolution_clock >	start_time

std::shared_ptr< NoAdaptivity< VectorType > >	adaptor_default

std::shared_ptr< DataOutput< dim, VectorType > >	data_out_default

double	Lambda

int	verbosity

double	output_dt

struct DiFfRG::AbstractTimestepper::ImplicitParameters	impl

struct DiFfRG::AbstractTimestepper::ExplicitParameters	expl

std::unordered_map< std::string, ConsoleOutStats >	console_out_stats_by_category

Static Protected Attributes
static constexpr uint	dim = dim_

static constexpr size_t	console_name_width = 21

Detailed Description

template<typename VectorType_, typename SparseMatrixType_, uint dim_>
class DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >

The abstract base class for all timestepping algorithms. It provides a standard constructor which populates typical timestepping parameters from a given JSONValue object, such as the timestep sizes, tolerances, verbosity, etc. that are used in the timestepping algorithms.

In the JSONValue object, a /timestepping/ section must be present with the following parameters:

/timestepping/output_dt: The output timestep size.
/timestepping/implicit/dt: The timestep size for an implicit timestepping algorithm.
/timestepping/implicit/minimal_dt: The minimal timestep size for an implicit timestepping algorithm.
/timestepping/implicit/maximal_dt: The maximal timestep size for an implicit timestepping algorithm.
/timestepping/implicit/abs_tol: The absolute tolerance for an implicit timestepping algorithm.
/timestepping/implicit/rel_tol: The relative tolerance for an implicit timestepping algorithm.
/timestepping/implicit/max_steps: The maximal number of internal SUNDIALS steps between outputs.
/timestepping/explicit/dt: The timestep size for an explicit timestepping algorithm.
/timestepping/explicit/minimal_dt: The minimal timestep size for an explicit timestepping algorithm.
/timestepping/explicit/maximal_dt: The maximal timestep size for an explicit timestepping algorithm.
/timestepping/explicit/abs_tol: The absolute tolerance for an explicit timestepping algorithm.
/timestepping/explicit/rel_tol: The relative tolerance for an explicit timestepping algorithm.

Additionally, the following parameters are being used:

/output/verbosity: The verbosity level of the output.
/physical/Lambda: The RG scale parameter Lambda. If not present, no RG scale is given when console_out is called.

The console_out method is used to print information about the current time, the calculation time and the current RG scale k to the console in a standardized way.

Template Parameters

VectorType_	The type of the vector used in the timestepping algorithm.. Currently only Vector<double> is supported.
SparseMatrixType_	The type of the sparse matrix used in the timestepping algorithm. This depends on the assembler used in the computation.
dim_	The dimensionality of the spatial discretization.

Member Typedef Documentation

◆ BlockVectorType

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

using DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::BlockVectorType = dealii::BlockVector<NumberType>

protected

◆ InverseSparseMatrixType

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

using DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::InverseSparseMatrixType = typename get_type::InverseSparseMatrixType<SparseMatrixType>

protected

◆ NumberType

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

using DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::NumberType = typename get_type::NumberType<VectorType>

protected

◆ SparseMatrixType

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

using DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::SparseMatrixType = SparseMatrixType_

protected

◆ VectorType

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

using DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::VectorType = VectorType_

protected

Constructor & Destructor Documentation

◆ AbstractTimestepper()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::AbstractTimestepper	(	const JSONValue &	json,
		AbstractAssembler< VectorType, SparseMatrixType, dim > *	assembler,
		DataOutput< dim, VectorType > *	data_out = `nullptr`,
		AbstractAdaptor< VectorType > *	adaptor = `nullptr`
	)

inline

Construct a new Abstract Timestepper object.

Parameters

json	The JSONValue object must contain a /timestepping/ section with all necessary parameters.
assembler
data_out
adaptor

Member Function Documentation

◆ console_out()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

void DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::console_out	(	const double	t,
		const std::string	name,
		const int	verbosity_level,
		const double	calc_dt_ms = `-1.0`
	)		const

inlineprotected

Pretty-print the status of the timestepping algorithm to the console.

Parameters

t	Current time.
name	A tag prepended to the output.
verbosity_level	The verbosity level of the output.
calc_dt_ms	If >= 0, the time in milliseconds it took to calculate the current timestep.

◆ console_out_category()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

static std::string DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::console_out_category ( const std::string & name )

inlinestaticprotected

◆ format_uncertain_ms()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

static std::string DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::format_uncertain_ms	(	const double	mean_ms,
		const double	uncertainty_ms
	)

inlinestaticprotected

◆ get_adaptor()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

AbstractAdaptor< VectorType > * DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::get_adaptor ( )

inline

Utility function to obtain an Adaptor object. If no Adaptor object is provided, a default one is created, which is the NoAdaptivity object, i.e. no mesh adaptivity is used.

Returns: AbstractAdaptor<VectorType>* A pointer to the Adaptor object.

◆ get_data_out()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

DataOutput< dim, VectorType > * DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::get_data_out ( )

inline

Utility function to obtain a DataOutput object. If no DataOutput object is provided, a default one is created.

Returns: DataOutput<dim, VectorType>* A pointer to the DataOutput object.

◆ print_console_out()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

void DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::print_console_out	(	const std::string &	name,
		const double	t,
		const size_t	milliseconds,
		const std::optional< std::string >	calc_dt = `std::nullopt`,
		const size_t	calls = `0`
	)		const

inlineprotected

◆ run()

template<typename VectorType_ , typename SparseMatrixType_ , uint dim_>

virtual void DiFfRG::AbstractTimestepper< VectorType_, SparseMatrixType_, dim_ >::run	(	AbstractFlowingVariables< NumberType > *	initial_condition,
		const double	t_start,
		const double	t_stop
	)

pure virtual

Any derived class must implement this method to run the timestepping algorithm.

Parameters

initial_condition	A pointer to a flowing variables object that contains the initial condition.
t_start	The start time of the simulation.
t_stop	The run method will evolve the system from t_start to t_stop.

Implemented in DiFfRG::TimeStepperBoostRK< VectorType, SparseMatrixType, dim, prec >, DiFfRG::TimeStepperImplicitEuler< VectorType, SparseMatrixType, dim, LinearSolver >, DiFfRG::TimeStepperSUNDIALS_IDA< VectorType, SparseMatrixType, dim, LinearSolver >, DiFfRG::TimeStepperSUNDIALS_IDA_BoostABM< VectorType, SparseMatrixType, dim, LinearSolver >, DiFfRG::TimeStepperSUNDIALS_IDA_BoostRK< VectorType, SparseMatrixType, dim, LinearSolver, prec >, and DiFfRG::TimeStepperTRBDF2< VectorType, SparseMatrixType, dim, LinearSolver >.