File: README

package info (click to toggle)
sparskit 2.0.0-4
  • links: PTS, VCS
  • area: main
  • in suites: bookworm, bullseye, sid
  • size: 4,348 kB
  • sloc: fortran: 15,253; makefile: 260; sh: 136; ansic: 76
file content (57 lines) | stat: -rw-r--r-- 2,475 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
-------------------------------------------------------------
		SPARSKIT MODULE MATGEN 
-------------------------------------------------------------

 The current directory MATGEN contains a few subroutines and
 drivers for generating sparse matrices.

      1) 5-pt and 7-pt matrices on rectangular regions discretizing
           elliptic operators of the form: 

         L u == delx( a delx u ) + dely ( b dely u) + delz ( c delz u ) + 
                delx ( d u ) + dely (e u) + delz( f u ) + g u = h u

         with Boundary conditions,
		 alpha del u / del n + beta u = gamma
	 on a rectangular 1-D, 2-D or 3-D grid using centered
	 difference scheme or upwind scheme.
 
         The functions a, b, ..., h are known through the
         subroutines  afun, bfun, ..., hfun in the file
         functns.f. The alpha is a constant on each side of the
	 rectanglar domain. the beta and the gamma are defined
	 by the functions betfun and gamfun (see functns.f for
	 examples).

      2) block version of the finite difference matrices (several degrees of
         freedom per grid point. ) It only generates the matrix (without
	 the right-hand-side), only Dirichlet Boundary conditions are used.

      3) Finite element matrices for the convection-diffusion problem 

                  - Div ( K(x,y) Grad u ) + C(x,y) Grad u = f
                    u = 0 on boundary 

	 (with Dirichlet boundary conditions). The matrix is returned 
	 assembled in compressed sparse row format. See genfeu for 
	 matrices in unassembled form. The user must provide the grid, 
	 (coordinates x, y and connectivity matrix ijk) as well as some 
	 information on the nodes (nodcode) and the material properties 
	 (the function K(x,y) above) in the form of a subroutine xyk. 

     4) Markov chain matrices arising from a random walk on a
	trangular grid. Useful for testing nonsymmetric eigenvalue
	codes. Has been suggested by G.W. Stewart in one of his
	papers. Used by Y. Saad in several papers as a test problem 
	for nonsymmetric eigenvalue methods.

     5) Matrices from the paper by Z. Zlatev, K. Schaumburg, 
        and J. Wasniewski. (``A testing scheme for subroutines solving
        large linear problems.''  Computers and Chemistry, 5:91--100, 
        1981.)

----------------------------------------------------------------------
     the items (1) and (2) are in directory FDIF,
     the item (3) is in directory FEM
     the items (4) and (5) are in directory MISC