[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: chol problem in 2.9.5
From: |
John W. Eaton |
Subject: |
Re: chol problem in 2.9.5 |
Date: |
Wed, 31 May 2006 22:32:02 -0400 |
On 31-May-2006, Dmitri A. Sergatskov wrote:
| On 5/26/06, John W. Eaton <address@hidden> wrote:
| ...
| > | octave:3> c=chol(b)
| > | error: chol: matrix not positive definite
| > | error: evaluating assignment expression near line 3, column 2
| >
| > I'm unable to duplicate this problem with my current sources.
| >
|
| John,
| Did you compile with --enable-64=yes?
No.
| If I recompile src.rpm with --enable-64=no the problem goes away...
| (this is with gcc 4.1.1)
OK, when you compile with --enable-64, what LAPACK and BLAS libraries
are you linking with? Are they compiled in a way that makes integers
8 bytes wide?
jwe
- Re: chol problem in 2.9.5, (continued)
- Re: chol problem in 2.9.5, Quentin Spencer, 2006/05/25
- Re: chol problem in 2.9.5, David Bateman, 2006/05/25
- Re: chol problem in 2.9.5, Quentin Spencer, 2006/05/25
- Re: chol problem in 2.9.5, Przemek Klosowski, 2006/05/26
- Re: chol problem in 2.9.5, Dmitri A. Sergatskov, 2006/05/26
- Re: chol problem in 2.9.5, John W. Eaton, 2006/05/26
- Re: chol problem in 2.9.5, Dmitri A. Sergatskov, 2006/05/26
- Re: chol problem in 2.9.5, Quentin Spencer, 2006/05/26
- Re: chol problem in 2.9.5, Dmitri A. Sergatskov, 2006/05/26
- Re: chol problem in 2.9.5, Dmitri A. Sergatskov, 2006/05/31
- Re: chol problem in 2.9.5,
John W. Eaton <=