1
function [x, info] = klus (A, b)
2
% [x, Info] = klus (A, b)
4
% Info (1): # of blocks
5
% Info (2): dimension of largest block
6
% Info (3): nnz (L+U+off diagonal blocks)
9
% usage: [x, Info] = klus (A, b, tol)
11
% tol: partial pivoting tolerance. Use 1e-3 (for example) to prefer diagonal
14
% THis is out of date:
16
% Info (2): nz in off diagonal part
17
% Info (3): # of blocks
18
% Info (4): max nz in diagonal blocks of A
19
% Info (5): dimension of largest block
20
% Info (6): estimated nz in L, incl. diagonal, excludes off-diagonal entries
21
% Info (7): estimated nz in U, incl. diagonal, excludes off-diagonal entries
23
% Info (8): nz in L, including diagonal, excludes off-diagonal entries
24
% Info (9): nz in U, including diagonal, excludes off-diagonal entries
25
% Info (10): analyze cputime
26
% Info (11): factor cputime
27
% Info (12): solve cputime
28
% Info (13): refactorize cputime (if computed)
29
% Info (14): # off-diagonal pivots chosen
31
% b may be n-by-m with m > 1. It must be dense.
35
error ('klus mexFunction not found') ;