reordersequence.h File Reference

Go to the source code of this file.

Functions
void	compute_sequence (const struct UnstructuredGrid *grid, const double *flux, int *sequence, int *components, int *ncomponents)
void	compute_sequence_graph (const struct UnstructuredGrid *grid, const double *flux, int *sequence, int *components, int *ncomponents, int *ia, int *ja)

Detailed Description

Facilities for computing a causal permutation of the cells in an UnstructuredGrid such that fluid transport may be subsequently solved by going from sources and up-stream cells to sinks and down-stream cells.

Function Documentation

compute_sequence()

void compute_sequence	(	const struct UnstructuredGrid *	grid,
		const double *	flux,
		int *	sequence,
		int *	components,
		int *	ncomponents
	)

Compute causal permutation sequence of grid cells with respect to specific Darcy flux field.

Parameters

[in]	grid	Grid structure for which to compute causal cell permutation.
[in]	flux	Darcy flux field. One scalar value for each interface/connection in the grid, including the boundary. We assume that flux[f] is positive if the flow is from cell grid->face_cells[2*f + 0] to cell grid->face_cells[2*f + 1].
[out]	sequence	Causal grid cell permutation. Ordered according to topological sorting of the strongly connected components of the Darcy flux upwind graph. Array of size grid->number_of_cells.
[out]	components	Indirection pointers that describe the strongly connected components (i.e., loops) of the Darcy flux upwind graph. Specifically, the 'th strongly connected component constitutes cells sequence[components[i] ... components[i + 1] 1].

In the ideal case of a perfectly reordered upwind graph, this array will hold grid->number_of_cells + 1 elements, and the relation components[i + 1] - components[i] == 1 holds for all . Thus, the components parameter must point to an array of grid->number_of_cells + 1 elements.

Parameters

[out]

ncomponents

Number of strongly connected components. Pointer to a single integer. The components output are valid for all i = 0 ... *ncomponents - 1. Furthermore, the number of components must be in the interval [1 .. grid->number_of_cells].

compute_sequence_graph()

void compute_sequence_graph	(	const struct UnstructuredGrid *	grid,
		const double *	flux,
		int *	sequence,
		int *	components,
		int *	ncomponents,
		int *	ia,
		int *	ja
	)

Compute causal permutation sequence of grid cells with respect to specific Darcy flux field. Also return the permuted upwind graph.

Parameters

[in]	grid	Grid structure for which to compute causal cell permutation.
[in]	flux	Darcy flux field. One scalar value for each interface/connection in the grid, including the boundary. We assume that flux[f] is positive if the flow is from cell grid->face_cells[2*f + 0] to cell grid->face_cells[2*f + 1].
[out]	sequence	Causal grid cell permutation. Ordered according to topological sorting of the strongly connected components of the Darcy flux upwind graph. Array of size grid->number_of_cells.
[out]	components	Indirection pointers that describe the strongly connected components (i.e., loops) of the Darcy flux upwind graph. Specifically, the 'th strongly connected component constitutes cells sequence[components[i] ... components[i + 1] 1].

In the ideal case of a perfectly reordered upwind graph, this array will hold grid->number_of_cells + 1 elements, and the relation components[i + 1] - components[i] == 1 holds for all . Thus, the components parameter must point to an array of grid->number_of_cells + 1 elements.

Parameters

[out]	ncomponents	Number of strongly connected components. Pointer to a single integer. The components output are valid for all i = 0 ... *ncomponents - 1. Furthermore, the number of components must be in the interval [1 .. grid->number_of_cells].
[out]	ia	Indirection pointers into ja that define the "upwind" cells that influence a particular grid cell.
[out]	ja	Compressed-sparse representation of the upwind graph. Specifically, the upwind cells that influence cell are ja[ia[i] .. ia[i+1]-1]. Array of size at least equal to the number of internal faces of grid. The number grid->number_of_faces is an upper bound of the array size.