LinearInterpolator3D< NT, Coordinates > Class Template Reference#

DiFfRG: DiFfRG::LinearInterpolator3D< NT, Coordinates > Class Template Reference

DiFfRG

Discretization Framework for functional Renormalization Group flows

A linear interpolator for 3D data, callable from host AND device code. More...

#include <linear_interpolator_3d.hh>

Public Types
using	ctype = typename Coordinates::ctype

using	value_type = NT

Public Member Functions
	LinearInterpolator3D (const Coordinates &coordinates)
	Construct a LinearInterpolator3D with internal, zeroed data and a coordinate system.

KOKKOS_DEFAULTED_FUNCTION	LinearInterpolator3D (const LinearInterpolator3D &)=default
	Shallow copy of BOTH views, valid in host and in device code. See LinearInterpolator1D.

template<typename NT2 >
void	update (const NT2 *in_data)
	Replace the data, leaving host AND device current. The only mutator.

NT	operator[] (size_t i) const
	Host-side element access, in the row-major order update() takes its input in.

device::array< ctype, 3 > KOKKOS_FUNCTION	index (const ctype &x, const ctype &y, const ctype &z) const
	Map physical coordinates onto grid indices.

NT KOKKOS_FUNCTION	at (const device::array< ctype, 3 > &idx) const
	Interpolate at grid indices previously obtained from index().

NT KOKKOS_FUNCTION	operator() (const ctype &x, const ctype &y, const ctype &z) const
	Interpolate the data at a given point.

const Coordinates &	get_coordinates () const
	Get the coordinate system of the data.

const NT *	data () const
	Read-only handle to the host values, in the mirror's storage order.

Static Public Attributes
static constexpr size_t	dim = 3

Private Types
using	ViewType = Kokkos::View<NT ***, GPU_memory, Kokkos::MemoryTraits<Kokkos::RandomAccess>>

using	HostViewType = typename ViewType::host_mirror_type

Private Member Functions
KOKKOS_FORCEINLINE_FUNCTION NT	value (const size_t i, const size_t j, const size_t k) const
	Read one element from whichever buffer belongs to the executing side.

Private Attributes
const Coordinates	coordinates

const device::array< size_t, 3 >	sizes

ViewType	device_data

HostViewType	host_data

Static Private Attributes
static constexpr bool	periodic_x = is_periodic_axis_v<Coordinates, 0>

static constexpr bool	periodic_y = is_periodic_axis_v<Coordinates, 1>

static constexpr bool	periodic_z = is_periodic_axis_v<Coordinates, 2>

static constexpr bool	has_separate_device

Detailed Description

template<typename NT, typename Coordinates>
class DiFfRG::LinearInterpolator3D< NT, Coordinates >

A linear interpolator for 3D data, callable from host AND device code.

See LinearInterpolator1D for the host/device dispatch rationale.

Template Parameters

NT	input data type
Coordinates	coordinate system of the input data

Member Typedef Documentation

◆ ctype

template<typename NT , typename Coordinates >

using DiFfRG::LinearInterpolator3D< NT, Coordinates >::ctype = typename Coordinates::ctype

◆ HostViewType

template<typename NT , typename Coordinates >

using DiFfRG::LinearInterpolator3D< NT, Coordinates >::HostViewType = typename ViewType::host_mirror_type

private

◆ value_type

template<typename NT , typename Coordinates >

using DiFfRG::LinearInterpolator3D< NT, Coordinates >::value_type = NT

◆ ViewType

template<typename NT , typename Coordinates >

using DiFfRG::LinearInterpolator3D< NT, Coordinates >::ViewType = Kokkos::View<NT ***, GPU_memory, Kokkos::MemoryTraits<Kokkos::RandomAccess>>

private

Constructor & Destructor Documentation

◆ LinearInterpolator3D() [1/2]

template<typename NT , typename Coordinates >

DiFfRG::LinearInterpolator3D< NT, Coordinates >::LinearInterpolator3D ( const Coordinates & coordinates )

inline

Construct a LinearInterpolator3D with internal, zeroed data and a coordinate system.

Parameters

coordinates coordinate system of the data

◆ LinearInterpolator3D() [2/2]

template<typename NT , typename Coordinates >

KOKKOS_DEFAULTED_FUNCTION DiFfRG::LinearInterpolator3D< NT, Coordinates >::LinearInterpolator3D ( const LinearInterpolator3D< NT, Coordinates > & )

default

Shallow copy of BOTH views, valid in host and in device code. See LinearInterpolator1D.

Member Function Documentation

◆ at()

template<typename NT , typename Coordinates >

NT KOKKOS_FUNCTION DiFfRG::LinearInterpolator3D< NT, Coordinates >::at ( const device::array< ctype, 3 > & idx ) const

inline

Interpolate at grid indices previously obtained from index().

◆ data()

template<typename NT , typename Coordinates >

const NT * DiFfRG::LinearInterpolator3D< NT, Coordinates >::data ( ) const

inline

Read-only handle to the host values, in the mirror's storage order.

NOTE the storage order is NOT the row-major order update() takes; see LinearInterpolator2D.

◆ get_coordinates()

template<typename NT , typename Coordinates >

const Coordinates & DiFfRG::LinearInterpolator3D< NT, Coordinates >::get_coordinates ( ) const

inline

Get the coordinate system of the data.

Returns: const Coordinates& the coordinate system

◆ index()

template<typename NT , typename Coordinates >

device::array< ctype, 3 > KOKKOS_FUNCTION DiFfRG::LinearInterpolator3D< NT, Coordinates >::index	(	const ctype &	x,
		const ctype &	y,
		const ctype &	z ) const

inline

Map physical coordinates onto grid indices.

Split out of operator() because it is the expensive half: Coordinates::backward is a fp64 log/log1p per logarithmic axis, ~200 fp64 instructions each on current NVIDIA parts against ~10 for the interpolation. A generated kernel evaluating several dressings at the SAME point otherwise pays it once per dressing – the compiler cannot CSE it, because each interpolator owns its own coordinates members and cannot be proven to agree with another's.

Depends only on the coordinate system, so the result may be shared across interpolators that share one. Clamping and stencil resolution stay in at(), since they are size dependent and would otherwise make an index untransferable between interpolators of different extent.

The return type is spelled out rather than deduced: nvcc loses decltype(var) when the initializer has a deduced return type inside a class-template member, and the consumer stores this in a const auto.

◆ operator()()

template<typename NT , typename Coordinates >

NT KOKKOS_FUNCTION DiFfRG::LinearInterpolator3D< NT, Coordinates >::operator()	(	const ctype &	x,
		const ctype &	y,
		const ctype &	z ) const

inline

Interpolate the data at a given point.

◆ operator[]()

template<typename NT , typename Coordinates >

NT DiFfRG::LinearInterpolator3D< NT, Coordinates >::operator[] ( size_t i ) const

inline

Host-side element access, in the row-major order update() takes its input in.

◆ update()

template<typename NT , typename Coordinates >

template<typename NT2 >

void DiFfRG::LinearInterpolator3D< NT, Coordinates >::update ( const NT2 * in_data )

inline

Replace the data, leaving host AND device current. The only mutator.

in_data is row-major; see LinearInterpolator2D::update() for why the fill indexes the mirror through operator() instead of copying flat, and LinearInterpolator1D::update() for why the single trailing fence is not optional.

◆ value()

template<typename NT , typename Coordinates >

KOKKOS_FORCEINLINE_FUNCTION NT DiFfRG::LinearInterpolator3D< NT, Coordinates >::value	(	const size_t	i,
		const size_t	j,
		const size_t	k ) const

inlineprivate

Read one element from whichever buffer belongs to the executing side.

Member Data Documentation

◆ coordinates

template<typename NT , typename Coordinates >

const Coordinates DiFfRG::LinearInterpolator3D< NT, Coordinates >::coordinates

private

◆ device_data

template<typename NT , typename Coordinates >

ViewType DiFfRG::LinearInterpolator3D< NT, Coordinates >::device_data

private

◆ dim

template<typename NT , typename Coordinates >

size_t DiFfRG::LinearInterpolator3D< NT, Coordinates >::dim = 3

staticconstexpr

◆ has_separate_device

template<typename NT , typename Coordinates >

bool DiFfRG::LinearInterpolator3D< NT, Coordinates >::has_separate_device

staticconstexprprivate

Initial value:

=

!std::is_same_v<typename ViewType::memory_space, typename HostViewType::memory_space>

◆ host_data

template<typename NT , typename Coordinates >

HostViewType DiFfRG::LinearInterpolator3D< NT, Coordinates >::host_data

private

◆ periodic_x

template<typename NT , typename Coordinates >

bool DiFfRG::LinearInterpolator3D< NT, Coordinates >::periodic_x = is_periodic_axis_v<Coordinates, 0>

staticconstexprprivate

◆ periodic_y

template<typename NT , typename Coordinates >

bool DiFfRG::LinearInterpolator3D< NT, Coordinates >::periodic_y = is_periodic_axis_v<Coordinates, 1>

staticconstexprprivate

◆ periodic_z

template<typename NT , typename Coordinates >

bool DiFfRG::LinearInterpolator3D< NT, Coordinates >::periodic_z = is_periodic_axis_v<Coordinates, 2>

staticconstexprprivate

◆ sizes

template<typename NT , typename Coordinates >

const device::array<size_t, 3> DiFfRG::LinearInterpolator3D< NT, Coordinates >::sizes

private

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

/home/runner/work/DiFfRG_current/DiFfRG_current/DiFfRG/include/DiFfRG/physics/interpolation/linear_interpolator_3d.hh

LinearInterpolator3D&lt; NT, Coordinates &gt; Class Template Reference#

Public Types

Public Member Functions

Static Public Attributes

Private Types

Private Member Functions

Private Attributes

Static Private Attributes

Detailed Description

Member Typedef Documentation

◆ ctype

◆ HostViewType

◆ value_type

◆ ViewType

Constructor & Destructor Documentation

◆ LinearInterpolator3D() [1/2]

◆ LinearInterpolator3D() [2/2]

Member Function Documentation

◆ at()

◆ data()

◆ get_coordinates()

◆ index()

◆ operator()()

◆ operator[]()

◆ update()

◆ value()

Member Data Documentation

◆ coordinates

◆ device_data

◆ dim

◆ has_separate_device

◆ host_data

◆ periodic_x

◆ periodic_y

◆ periodic_z

◆ sizes

LinearInterpolator3D< NT, Coordinates > Class Template Reference#