Detailed Description

Routines to compute the pattern of the result of a backsolve.

Author: Ben Hermans

These routines are used to compute the pattern of the (next) row in L during the numerical factorization, or more generally as a first step in performing a backsolve operation with a sparse right hand side.

Definition in file ladel_pattern.h.

#include "ladel_types.h"

Include dependency graph for ladel_pattern.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
ladel_int	ladel_nonzero_pattern_of_row_in_L (ladel_sparse_matrix M, ladel_symbolics sym, ladel_int row)
	Computes the pattern of the (next) row in L. More...

ladel_int	ladel_etree_dfs (ladel_sparse_matrix W, ladel_symbolics sym, ladel_int col_in_W, ladel_int maximum_row)
	Computes a depth-first search of the given pattern through the elimination tree. More...

Function Documentation

◆ ladel_etree_dfs()

ladel_int ladel_etree_dfs	(	ladel_sparse_matrix *	W,
		ladel_symbolics *	sym,
		ladel_int	col_in_W,
		ladel_int	maximum_row
	)

Computes a depth-first search of the given pattern through the elimination tree.

This routine computes the pattern of the result of a backsolve (with a sparse right hand side).

Note: This pattern is stored in sym->pattern[start] through sym->pattern[m-1], where start is the return value of this function.

Parameters

W	Right hand side vector of the backsolve
sym	Symbolics struct
col_in_W	Column in W to consider
maximum_row	Perform the Gaussian elimination only up until this row (not included)

Returns: Starting index for the pattern

Definition at line 40 of file ladel_pattern.c.

Here is the caller graph for this function:

◆ ladel_nonzero_pattern_of_row_in_L()

ladel_int ladel_nonzero_pattern_of_row_in_L	(	ladel_sparse_matrix *	M,
		ladel_symbolics *	sym,
		ladel_int	row
	)

Computes the pattern of the (next) row in L.

Note: This pattern is stored in sym->pattern[start] through sym->pattern[m-1], where start is the return value of this function.