help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Inverse Matrix Function appears a bit wonky


From: John W. Eaton
Subject: Re: Inverse Matrix Function appears a bit wonky
Date: Thu, 4 Nov 2004 11:41:41 -0500

On  4-Nov-2004, David Bateman <address@hidden> wrote:

| According to Robert A. Macy <address@hidden> (on 11/04/04):
| > That's it!  simply asking to return the value of rcond
| > allows me to catch a "bad" matrix.  
| 
| If you are after speed don't just ask for the condition number everytime.
| If you have a good expectation that the matrix isn't singular, not 
| calculating the condition number is a speed gain.

The condition number is always calculated by the underlying LAPACK
routines, so I'm not sure that there would be a significant speed
difference.

There were a couple of problems that had been introduced, probably
when changes were made to inv and det to return the condition number.
The critical one was that when deciding whether to print the warning
message, the status returned from the LAPACK routines was checked, but
not the value of rcond.  Now rcond is checked (I hope in a way that
will not fail due to any compiler optimization level) so I think the
problem is now fixed in my sources.

Thanks,

jwe



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------



reply via email to

[Prev in Thread] Current Thread [Next in Thread]