type_int_point_1d Module Reference

The type for generic 1D point interpolation that has to be derived.

Data Types
interface	int_point_1D_init
	Abstract interfaces. More...

Functions/Subroutines
subroutine	int_point_1d_construct (self, index_start, index_end)
	Initialize a 1D point interpolant. More...
subroutine	int_point_1d_place (self, X, coordinates, control_volumes)
	Place (position) the interpolant to the corresponding 'index' given the coordinate 'X' of the point and the discrete nodes. The resulting index corresponds to the node that is just before (less than or equal) the point. More...
double precision function	int_point_1d_apply (self, values)
	Apply the stencil of the interpolant after it has been placed, given the array of discrete node values. The result is the interpolation. More...
double precision function	int_point_1d_interpolate (self, X, coordinates, values, derivative, control_volumes)
	Directly compute the interpolation at the given position. The result is the interpolation value. More...
pure integer function	compute_point_index_sorted_array_dichotomic (X, coordinates)
	Returns the index of the node that is just before (<=) 'X'. More...

Function/Subroutine Documentation

compute_point_index_sorted_array_dichotomic()

pure integer function type_int_point_1d::compute_point_index_sorted_array_dichotomic	(	double precision, intent(in)	X,
		double precision, dimension(:), intent(in)	coordinates
	)

Parameters

[in]	X	the point position
[in]	coordinates	the coordinates of the nodes (in the interval '1:nx')

Returns

the resulting node index

Note

the algorithm is dichotomic and should be the most efficient in the general "random X" case

int_point_1d_apply()

double precision function type_int_point_1d::int_point_1d_apply	(	class(t_int_point_1d), intent(in)	self,
		double precision, dimension(:), intent(in)	values
	)

Parameters

[in]	self	the element
[in]	values	the values in the interval '1:nx'

Precondition

- the subroutine 'place' has been previously called

- the subroutine 'compute_stencil' have been previously called

int_point_1d_construct()

subroutine type_int_point_1d::int_point_1d_construct	(	class(t_int_point_1d), intent(inout)	self,
		integer, intent(in)	index_start,
		integer, intent(in)	index_end
	)

Todo:

MCO: bring back the "construct" function instead for more clarity

int_point_1d_interpolate()

double precision function type_int_point_1d::int_point_1d_interpolate	(	class(t_int_point_1d), intent(inout)	self,
		double precision, intent(in)	X,
		double precision, dimension(:), intent(in)	coordinates,
		double precision, dimension(:), intent(in)	values,
		integer, intent(in), optional	derivative,
		double precision, dimension(:), intent(in), optional	control_volumes
	)

Parameters

[in,out]	self	the element
[in]	X	the position where to interpolate
[in]	coordinates	the coordinates of the nodes (in the interval '1:nx')
[in]	values	the values (in the interval '1:nx')
[in]	control_volumes	the unused control volumes (used only for reconstruction)

int_point_1d_place()

subroutine type_int_point_1d::int_point_1d_place	(	class(t_int_point_1d), intent(inout)	self,
		double precision, intent(in)	X,
		double precision, dimension(:), intent(in)	coordinates,
		double precision, dimension(:), intent(in), optional	control_volumes
	)

Warning

if 'X' is less than 'coordinates(1)', then 'index' is set to '1'.

if 'X' is greater than 'coordinates(nx)', then 'index' is set to 'nx'.

Note

the index search algorithm is dichotomic

Precondition

coordinates is ordered and is at least of size 2