home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #20 / NN_1992_20.iso / spool / comp / sys / super / 931 next >
Encoding:
Text File  |  1992-09-07  |  1.9 KB  |  46 lines
  1. Newsgroups: comp.sys.super
  2. Path: sparky!uunet!sun-barr!ames!agate!spool.mu.edu!umn.edu!lynx!nmsu.edu!opus!jamartin
  3. From: jamartin@nmsu.edu (Jay Martinez)
  4. Subject: CMSSL routine for solving 2D finite difference?
  5. Message-ID: <JAMARTIN.92Sep5220655@gauss.nmsu.edu>
  6. Sender: usenet@nmsu.edu
  7. Organization: NMSU Computer Science
  8. Distribution: comp.sys.super
  9. Date: Sun, 6 Sep 1992 05:06:55 GMT
  10. Lines: 34
  11.  
  12.  
  13.  
  14.    I am trying to do a two-dimensional finite difference formulation
  15.    on a CM-200.  For the resulting system of equations Ax=B, I get an
  16.    A maxtrix that looks like this:
  17.  
  18.     
  19.                |  c    d              e                 |
  20.                |                                        |
  21.                |  b    c    d              e            |
  22.                |                                        |
  23.                |       b    c    d              e       |
  24.                |                                        |
  25.                |           b    c    d               e  |
  26.                |                                        |
  27.                |  a              b    c    d            |
  28.                |                                        |
  29.                |       a              b    c    d       |
  30.                |                                        |
  31.                |            a              b    c    d  |
  32.                |                                        |
  33.                |                 a             b    c   |
  34.  
  35.    where the space between the centre tridiagonal band and the outer
  36.    bands is two.  Does anyone know of a CMSSL routine that will solve
  37.    such a system?  Or is there anyway to transform this system into
  38.    one that can be solved by another CMSSL sparse system routine?
  39.    The CMSSL routine gen_penta_diag doesn't seem to have a provision
  40.    for spaces between bands.  I would appreciate any suggestions.
  41.  
  42.    Jay Jansen Martinez (jamartin@NMSU.Edu)
  43.    
  44.    Department of Electrical and Computer Engineering
  45.    New Mexico State University
  46.