detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v > Class Template Reference

Runge-Kutta-Nystrom 4th order stepper implementation. More...

#include <detray/propagator/rk_stepper.hpp>

Inheritance diagram for detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >:

Collaboration diagram for detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >:

Classes
struct	intermediate_state
	Intermediate values evaluated at the RK integration stages. More...
struct	state
	Per-track integration state maintained by the stepper. More...

Public Types
using	algebra_type = algebra_t
	Algebra backend used by this stepper.
using	bound_track_parameters_type
	Bound track parameter type accepted by the stepper.
using	free_track_parameters_type
	Free track parameter type advanced by the stepper.
using	magnetic_field_type = magnetic_field_t
	Magnetic field representation used by the stepper.
template<std::size_t ROWS, std::size_t COLS>
using	matrix_type = dmatrix<algebra_t, ROWS, COLS>
	Matrix type with the requested dimensions.
using	point3_type = dpoint3D<algebra_t>
	Three-dimensional point type.
using	scalar_type = dscalar<algebra_t>
	Scalar type associated with the algebra backend.
using	transform3_type = dtransform3D<algebra_t>
	Three-dimensional transform type.
using	transport_jacobian_type
	Transport Jacobian representation selected by the flags.
using	vector3_type = dvector3D<algebra_t>
	Three-dimensional vector type.

Public Member Functions
void	advance_jacobian (detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &stepping, const stepping::config &cfg, const intermediate_state &sd, const material< scalar_type > *vol_mat_ptr) const
	Update the jacobian transport from free propagation.
void	advance_track (detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &stepping, const intermediate_state &sd, const material< scalar_type > *vol_mat_ptr) const
	Update the track state by Runge-Kutta-Nystrom integration.
detray::pair< scalar_type, scalar_type >	evaluate_dqopds (detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &stepping, const std::size_t i, const scalar_type h, const scalar_type dqopds_prev, const material< scalar_type > *vol_mat_ptr, const detray::stepping::config &cfg) const
	evaluate dqopds for a given step size and material
detray::pair< vector3_type, vector3_type >	evaluate_dtds (detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &stepping, const vector3_type &b_field, const std::size_t i, const scalar_type h, const vector3_type &dtds_prev, const scalar_type qop) const
	Evaluate dtds for Runge-Kutta stepping.
matrix_type< 3, 3 >	evaluate_field_gradient (detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &stepping, const point3_type &pos) const
	Estimate the magnetic-field spatial gradient by centered differences.
bool	step (const scalar_type dist_to_next, state &stepping, const stepping::config &cfg, bool do_reset, const material< scalar_type > *vol_mat_ptr=nullptr) const
	Take a step, using an adaptive Runge-Kutta algorithm.

Static Public Attributes
static constexpr bool	uses_gradient
	True when field-gradient terms are available for transport.

Detailed Description

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>
class detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >

Runge-Kutta-Nystrom 4th order stepper implementation.

Template Parameters

magnetic_field_t	the type of magnetic field
algebra_t	algebra backend used for track calculations
constraint_t	constraints applied to the step size
policy_t	policy used to update navigation trust
inspector_t	optional stepping inspector
flags_v	bitmask of enabled stepper capabilities

Member Typedef Documentation

bound_track_parameters_type

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

using detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::bound_track_parameters_type

Initial value:

 
      typename base_type::bound_track_parameters_type

Bound track parameter type accepted by the stepper.

free_track_parameters_type

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

using detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::free_track_parameters_type

Initial value:

 
      typename base_type::free_track_parameters_type

Free track parameter type advanced by the stepper.

transport_jacobian_type

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

using detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::transport_jacobian_type

Initial value:

 std::conditional_t<
      uses_gradient,
      detail::transport_jacobian_matrix_with_gradient<algebra_type>,
      detail::transport_jacobian_matrix_without_gradient<algebra_type>>

Transport Jacobian representation selected by the flags.

Member Function Documentation

advance_jacobian()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

void detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::advance_jacobian	(	detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &	stepping,
		const stepping::config &	cfg,
		const intermediate_state &	sd,
		const material< scalar_type > *	vol_mat_ptr ) const

Update the jacobian transport from free propagation.

Parameters

stepping	State whose Jacobian is updated.
cfg	Stepping configuration controlling gradient terms.
sd	Evaluated RK stage data.
vol_mat_ptr	Material in the current volume, or null in vacuum.

advance_track()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

void detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::advance_track	(	detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &	stepping,
		const intermediate_state &	sd,
		const material< scalar_type > *	vol_mat_ptr ) const

Update the track state by Runge-Kutta-Nystrom integration.

Parameters

stepping	State to update.
sd	Evaluated RK stage data.
vol_mat_ptr	Material in the current volume, or null in vacuum.

evaluate_dqopds()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

detray::pair< scalar_type, scalar_type > detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::evaluate_dqopds	(	detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &	stepping,
		const std::size_t	i,
		const scalar_type	h,
		const scalar_type	dqopds_prev,
		const material< scalar_type > *	vol_mat_ptr,
		const detray::stepping::config &	cfg ) const

evaluate dqopds for a given step size and material

Parameters

stepping	State being integrated.
i	RK stage index.
h	Stage step offset.
dqopds_prev	Derivative from the preceding stage.
vol_mat_ptr	Material in the current volume, or null in vacuum.
cfg	Stepping configuration.

Returns

Derivative and charge-over-momentum value for this stage.

evaluate_dtds()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

detray::pair< vector3_type, vector3_type > detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::evaluate_dtds	(	detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &	stepping,
		const vector3_type &	b_field,
		const std::size_t	i,
		const scalar_type	h,
		const vector3_type &	dtds_prev,
		const scalar_type	qop ) const

Evaluate dtds for Runge-Kutta stepping.

Parameters

stepping	State being integrated.
b_field	Magnetic field at this stage.
i	RK stage index.
h	Stage step offset.
dtds_prev	Direction derivative from the preceding stage.
qop	Charge-over-momentum at this stage.

Returns

Direction derivative and tangential direction.

evaluate_field_gradient()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

matrix_type< 3, 3 > detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::evaluate_field_gradient	(	detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::state &	stepping,
		const point3_type &	pos ) const

Estimate the magnetic-field spatial gradient by centered differences.

Parameters

stepping	State providing the magnetic field.
pos	Position where the gradient is evaluated.

Returns

Matrix whose columns are field derivatives by coordinate.

step()

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

bool detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::step	(	const scalar_type	dist_to_next,
		state &	stepping,
		const stepping::config &	cfg,
		bool	do_reset,
		const material< scalar_type > *	vol_mat_ptr = nullptr ) const

Take a step, using an adaptive Runge-Kutta algorithm.

Parameters

dist_to_next	The straight line distance to the next surface
stepping	The state object of a stepper
cfg	The stepping configuration
do_reset	whether to reset the RKN step size to "dist to next"
vol_mat_ptr	Material in the current volume, or null in vacuum.

Returns

returning the heartbeat, indicating if the stepping is alive

Member Data Documentation

uses_gradient

template<typename magnetic_field_t, concepts::algebra algebra_t, typename constraint_t = unconstrained_step<dscalar<algebra_t>>, typename policy_t = stepper_rk_policy<dscalar<algebra_t>>, typename inspector_t = stepping::void_inspector, std::uint32_t flags_v = (static_cast<std::uint32_t>( rk_stepper_flags::e_allow_covariance_transport) | static_cast<std::uint32_t>( rk_stepper_flags::e_allow_field_gradient))>

bool detray::rk_stepper< magnetic_field_t, algebra_t, constraint_t, policy_t, inspector_t, flags_v >::uses_gradient

staticconstexpr

Initial value:

= static_cast<bool>(
      flags_v &
      static_cast<std::uint32_t>(rk_stepper_flags::e_allow_field_gradient))

True when field-gradient terms are available for transport.