#include <include/ryujin/typed/hyperbolic_module.h>

Inheritance diagram for ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >:

[legend]

Collaboration diagram for ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >:

[legend]

Public Member Functions
Constructor and setup
	HyperbolicModule (const MPIEnsemble &mpi_ensemble, const OfflineData< dim, Number > &offline_data, const HyperbolicSystem &hyperbolic_system, const InitialValues< HyperbolicDescription, dim, Number > &initial_values, const std::string &subsection="/HyperbolicModule")

void	prepare () override

Functions for performing explicit time steps
void	reinit_state_vector (StateVector< Number > &state_vector) const override

void	prepare_state_vector (StateVector< Number > &state_vector, Number t) const override

void	fill_precomputed_values (StateVector< Number > &state_vector) const override

Number	forward_euler_step (const StateVector< Number > &old_state_vector, StageVectors stage_state_vectors, StageWeights stage_weights, StateVector< Number > &new_state_vector, Number t, Number tau, Number tau_max=std::numeric_limits< Number >::max()) const override

Information and statistics
void	print_solver_statistics (std::ostream &) const override

Public Member Functions inherited from ryujin::Interface::HyperbolicModule< dim, Number >
void	set_cfl (Number new_cfl) const

void	set_acceptable_tau_max_ratio (Number new_acceptable_tau_max_ratio) const

Typedefs and constexpr constants
using	StageVectors = typename Interface::HyperbolicModule< dim, Number >::StageVectors

using	StageWeights = typename Interface::HyperbolicModule< dim, Number >::StageWeights

using	HyperbolicSystem = typename HyperbolicDescription::HyperbolicSystem

using	View = typename HyperbolicSystem::template View< dim, Number >

using	Indicator = typename HyperbolicDescription::template Indicator< Number >

using	Limiter = typename HyperbolicDescription::template Limiter< Number >

using	WaveSpeedEstimator = typename HyperbolicDescription::template WaveSpeedEstimator< Number >

using	state_type = typename View::state_type

using	precomputed_type = typename View::precomputed_type

using	initial_precomputed_type = typename View::initial_precomputed_type

static constexpr auto	problem_dimension = View::problem_dimension

static constexpr auto	n_precomputed_values = View::n_precomputed_values

static constexpr auto	n_initial_precomputed_values

Additional Inherited Members
Protected Member Functions inherited from ryujin::Interface::HyperbolicModule< dim, Number >
	HyperbolicModule (const unsigned int problem_dimension, const unsigned int n_precomputed_values, const std::string &subsection)

Protected Attributes inherited from ryujin::Interface::HyperbolicModule< dim, Number >
Number	cfl_

Number	acceptable_tau_max_ratio_

Detailed Description

template<typename HyperbolicDescription, int dim, typename Number = double>
class ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >

Explicit forward Euler time-stepping for hyperbolic systems with convex limiting.

This module is described in detail in [17], Alg. 1.

Definition at line 37 of file hyperbolic_module.h.

Member Typedef Documentation

◆ StageVectors

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::StageVectors = typename Interface::HyperbolicModule<dim, Number>::StageVectors

Definition at line 45 of file hyperbolic_module.h.

◆ StageWeights

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::StageWeights = typename Interface::HyperbolicModule<dim, Number>::StageWeights

Definition at line 48 of file hyperbolic_module.h.

◆ HyperbolicSystem

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::HyperbolicSystem = typename HyperbolicDescription::HyperbolicSystem

Definition at line 51 of file hyperbolic_module.h.

◆ View

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::View = typename HyperbolicSystem::template View<dim, Number>

Definition at line 53 of file hyperbolic_module.h.

◆ Indicator

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::Indicator = typename HyperbolicDescription::template Indicator<Number>

Definition at line 55 of file hyperbolic_module.h.

◆ Limiter

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::Limiter = typename HyperbolicDescription::template Limiter<Number>

Definition at line 58 of file hyperbolic_module.h.

◆ WaveSpeedEstimator

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::WaveSpeedEstimator = typename HyperbolicDescription::template WaveSpeedEstimator<Number>

Definition at line 60 of file hyperbolic_module.h.

◆ state_type

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::state_type = typename View::state_type

Definition at line 65 of file hyperbolic_module.h.

◆ precomputed_type

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::precomputed_type = typename View::precomputed_type

Definition at line 69 of file hyperbolic_module.h.

◆ initial_precomputed_type

template<typename HyperbolicDescription , int dim, typename Number = double>

using ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::initial_precomputed_type = typename View::initial_precomputed_type

Definition at line 74 of file hyperbolic_module.h.

Constructor & Destructor Documentation

◆ HyperbolicModule()

template<typename HyperbolicDescription , int dim, typename Number >

ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::HyperbolicModule	(	const MPIEnsemble &	mpi_ensemble,
		const OfflineData< dim, Number > &	offline_data,
		const HyperbolicSystem &	hyperbolic_system,
		const InitialValues< HyperbolicDescription, dim, Number > &	initial_values,
		const std::string &	subsection = `"/HyperbolicModule< HyperbolicDescription, dim, Number >"`
	)

Constructor

Definition at line 43 of file hyperbolic_module.template.h.

Member Function Documentation

◆ prepare()

template<typename HyperbolicDescription , int dim, typename Number >

void ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::prepare ( )

override

Prepare time stepping. A call to prepare() allocates temporary storage and is necessary before any of the following time-stepping functions can be called.

Definition at line 64 of file hyperbolic_module.template.h.

References ryujin::have_separate_memory_spaces, and ryujin::implicit_transfers.

◆ reinit_state_vector()

template<typename HyperbolicDescription , int dim, typename Number >

void ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::reinit_state_vector ( StateVector< Number > & state_vector ) const

override

(Re)initialize the hyperbolic state vector component and the precomputed vector component of the state vector.

Note: This routine does not modify the parabolic state vector component.

Helper function that (re)initializes the state and the precomputed component of a StateVector to proper sizes.

Note: This method does neither initialize nor resize the parabolic state vector component. The ParabolicModule itself has to ensure proper setup during prepare_state_vector() and solve().

Definition at line 158 of file hyperbolic_module.template.h.

◆ prepare_state_vector()

template<typename HyperbolicDescription , int dim, typename Number >

void ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::prepare_state_vector	(	StateVector< Number > &	state_vector,
		Number	t
	)		const

override

This function preprocesses a given state vector U in preparation for an explicit euler step performed by the forward_euler_step() function. The function performs the following tasks:

We update the U component of the state vector by enforcing boundary conditions for the supplied time time t. We then updates ghost ranges on U so that the state vector is consistent across MPI ranks.
The function then runs the precomputation loop that populates the "precomputed values" component of the state vector and distributes the result over all MPI ranks by updating ghost ranges of the precomputed values vector.

Definition at line 192 of file hyperbolic_module.template.h.

References ryujin::precomputed, and ryujin::state.

◆ fill_precomputed_values()

template<typename HyperbolicDescription , int dim, typename Number >

void ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::fill_precomputed_values ( StateVector< Number > & state_vector ) const

overridevirtual

Populate the precomputed values of the state vector state_vector from its hyperbolic state and update the ghost values. In contrast to prepare_state_vector() no boundary conditions are applied and constrained degrees of freedom are not skipped.

Note: This function operates on the selected memory space.

Implements ryujin::Interface::HyperbolicModule< dim, Number >.

Definition at line 232 of file hyperbolic_module.template.h.

References ryujin::StateVector< Number >::precomputed, ryujin::precomputed, and ryujin::state.

◆ forward_euler_step()

template<typename HyperbolicDescription , int dim, typename Number >

Number ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::forward_euler_step	(	const StateVector< Number > &	old_state_vector,
		StageVectors	stage_state_vectors,
		StageWeights	stage_weights,
		StateVector< Number > &	new_state_vector,
		Number	t,
		Number	tau,
		Number	tau_max = `std::numeric_limits<Number>::max()`
	)		const

overridevirtual

Given a reference to a previous state vector old_state_vector at time t perform an explicit euler step (and store the result in new_state_vector). The function returns the time step size tau with which the update was performed.

If tau is nonzero the step is performed with tau. If tau is 0 the step is performed with the minimum of tau_max and the computed maximal time step size according to the CFL condition.

The function takes an optional list of states stage_state_vectors together with a list of weights stage_weights to construct a modified high-order flux. The standard high-order flux reads (cf. [17], Eq. 12):

\begin{align} \newcommand{\bF}{{\boldsymbol F}} \newcommand{\bU}{{\boldsymbol U}} \newcommand\bUni{\bU^n_i} \newcommand\bUnj{\bU^n_j} \newcommand{\polf}{{\mathbb f}} \newcommand\Ii{\mathcal{I}(i)} \newcommand{\bc}{{\boldsymbol c}} \sum_{j\in\Ii} \frac{m_{ij}}{m_{j}} \; \frac{m_{j}}{\tau_n}\big( \tilde\bU_j^{H,n+1} - \bU_j^{n}\big) \;=\; \bF^n_i + \sum_{j\in\Ii}d_{ij}^{H,n}\big(\bUnj-\bUni\big), \qquad\text{with}\quad \bF^n_i\;:=\; \sum_{j\in\Ii}\Big(-(\polf(\bUni)+\polf(\bUnj)) \cdot\bc_{ij}\Big). \end{align}

Instead, the function replaces \(\bF^n_i\) in the right-hand side by the modified high-order flux:

\begin{align} \newcommand{\bF}{{\boldsymbol F}} \newcommand{\bU}{{\boldsymbol U}} \newcommand\bUnis{\bU^{s,n}_i} \newcommand\bUnjs{\bU^{s,n}_j} \newcommand{\polf}{{\mathbb f}} \newcommand\Ii{\mathcal{I}(i)} \newcommand{\bc}{{\boldsymbol c}} \tilde{\bF}^n_i\;:=\; \big(1-\sum_{s=1}^{\text{stages}}\omega_s\big)\bF^n_i \;+\; \sum_{s=1}^{\text{stages}}\omega_s \bF^{s,n}_i \qquad\text{with}\quad \bF^{s,n}_i\;:=\; \sum_{j\in\Ii}\Big(-(\polf(\bUnis)+\polf(\bUnjs)) \cdot\bc_{ij}\Big). \end{align}

where \(\omega_s\) denotes the weights for the given stages \(\bU^{s,n}\). The high-order viscosity term \(d_{ij}^{H,n}(\bU^n_j-\bU^n_i)\) is always evaluated with the old state. Source terms, if present, are combined with the same weights.

Note: The routine only performs an explicit update step on the locally owned dof index range. It neither updates the precomputed block of the state vector, sets boundary conditions (prior) to the update step, nor automatically updates the ghost range of the vector. It is thus necessary to call HyperbolicModule::prepare_state_vector() on old_state_vector prior to calling forward_euler_step().

Implements ryujin::Interface::HyperbolicModule< dim, Number >.

Definition at line 410 of file hyperbolic_module.template.h.

◆ print_solver_statistics()

template<typename HyperbolicDescription , int dim, typename Number = double>

void ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::print_solver_statistics ( std::ostream & ) const

inlineoverride

Print a status line with solver statistics. The explicit update has no solver statistics to report.

Definition at line 218 of file hyperbolic_module.h.

Member Data Documentation

◆ problem_dimension

template<typename HyperbolicDescription , int dim, typename Number = double>

constexpr auto ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::problem_dimension = View::problem_dimension

staticconstexpr

Definition at line 63 of file hyperbolic_module.h.

◆ n_precomputed_values

template<typename HyperbolicDescription , int dim, typename Number = double>

constexpr auto ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::n_precomputed_values = View::n_precomputed_values

staticconstexpr

Definition at line 67 of file hyperbolic_module.h.

◆ n_initial_precomputed_values

template<typename HyperbolicDescription , int dim, typename Number = double>

constexpr auto ryujin::HyperbolicModule< HyperbolicDescription, dim, Number >::n_initial_precomputed_values

staticconstexpr

Initial value:

=

View::n_initial_precomputed_values

Definition at line 71 of file hyperbolic_module.h.

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

include/ryujin/typed/hyperbolic_module.h
include/ryujin/typed/hyperbolic_module.template.h

Public Member Functions

Typedefs and constexpr constants

Additional Inherited Members

Detailed Description

Member Typedef Documentation

◆ StageVectors

◆ StageWeights

◆ HyperbolicSystem

◆ View

◆ Indicator

◆ Limiter

◆ WaveSpeedEstimator

◆ state_type

◆ precomputed_type

◆ initial_precomputed_type

Constructor & Destructor Documentation

◆ HyperbolicModule()

Member Function Documentation

◆ prepare()

◆ reinit_state_vector()

◆ prepare_state_vector()

◆ fill_precomputed_values()

◆ forward_euler_step()

◆ print_solver_statistics()

Member Data Documentation

◆ problem_dimension

◆ n_precomputed_values

◆ n_initial_precomputed_values