From bug-octave-request at bevo dot che dot wisc dot edu  Thu Jun  7 13:55:25 2001
Subject: Rcond=NAN in DET
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: "Jianming" <caijianming at yahoo dot co dot uk>
Cc: <bug-octave at bevo dot che dot wisc dot edu>
Date: Thu, 7 Jun 2001 13:55:20 -0500

On  7-Jun-2001, Jianming <caijianming at yahoo dot co dot uk> wrote:

| The following expression will cause octave 2.1.34 + atlas 3.2.1 to hang:
| 
| C=3D[1,  1.30007902878166e-112, 1,      1;
| 
|    9.2176458517918e-188, 1,   9.21764597429266e-188, =
| 9.21764617989136e-188;
| 
|    1,  1.30007903363547e-112, 1,       1;
| 
|    1,  1.30007904184461e-112, 1,       1];
| 
| 
| 
| [d r] =3D det(C);
| 
| 
| 
| This is because the DEGCO returns rcond =3D NaN, but was not tested by =
| dMatrix.cc. The method determinant thus proceed to perform DGEDI, which =
| will hang.
| 
| Attached is a patch which simply add the test condition xisnan(rcond).

OK, I made this change for Matrix::determinant and also the other
functions in dMatrix.cc and CMatrix.cc that call {d,z}geco.

I'm not sure this is the best solution, but it seems better than to
just go on after dgeco returns a NaN for rcond.

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
-------------------------------------------------------------