From help-request at octave dot org Sat Apr 15 16:32:49 2006 Subject: Re: chol? From: Marius Schamschula To: Joe Koski Cc: Vic Norton , Octave Help Date: Sat, 15 Apr 2006 16:31:16 -0500 --Apple-Mail-3--674248001 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Joe, Vic, Here is what I get on my Dual Core G5 under Mac OS X 10.4.6: octave:1> C = [36 24; 24 25]; octave:2> chol(C) ans = 6 4 0 3 Upper Triangular octave:3> version ans = 2.1.72 In other words, we got to fix this .oct problem. I just wish I knew where to start... On Apr 15, 2006, at 4:10 PM, Joe Koski wrote: > Vic, Marius, > > A further thought. chol is a .oct file. Are other .oct files > (trisolve.oct, etc.) working? It may be a dynamic load library > problem. The HPC version of octave was compiled on OS X 10.3.9, > and, as I recall, Vic is running 10.4.x. > > Joe > > > on 4/15/06 2:53 PM, Joe Koski at jkoski11 at comcast dot net wrote: > >> Vic, Marius, >> >> It must be the HPC version. On my home built with g95 version: >> >> octave:1> C = [36 24; 24 25]; >> octave:2> chol(C); >> octave:3> chol(C) >> ans = >> >> 6 4 >> 0 3 >> Upper Triangular >> >> octave:4> version >> ans = 2.1.73 >> >> >> Joe >> >> >> on 4/15/06 2:29 PM, Marius Schamschula at marius173 at mchsi dot com wrote: >> >>> Vic, >>> >>> Is this the port from hpc.sf.net, or did you compile from source? >>> >>> I have yet to get a working build of octave 2.1.73 on the Mac. >>> >>> On Apr 15, 2006, at 3:16 PM, Vic Norton wrote: >>> >>>> I had thought Octave, version 2.1.73, was working fine on my >>>> iMac G5, >>>> but now I'm not so sure. Here's my problem. >>>> >>>> I start with the matrix >>>> octave> C = [36 24; 24 25]; >>>> and do >>>> octave> chol(C); >>>> After waiting for a while I get >>>> panic: Bus error -- stopping myself... >>>> attempting to save variables to `octave-core'... >>>> save to `octave-core' complete >>>> Bus error >>>> vic$ >>>> >>>> To tell you the truth I'm not sure what the Cholesky factor of C is >>>> supposed to be. I do know that C = R' * R where >>>> R = [0 3; 6 4]. >>>> I have no problem computing another R that does this job: >>>> octave> C = [36 24; 24 25]; >>>> octave> [U, S, V] = svd(C); >>>> octave> R = sqrt(S) * V'; >>>> Now >>>> R = [ >>>> -5.8067 -4.6265 >>>> -1.5108 1.8962 >>>> ] >>>> and C = R' * R for this R as well. But R = chol(C) simply >>>> doesn't work >>>> for me. >>>> >>>> Is anybody else having any problems with "chol" or is it just my >>>> screwed >>>> up system? >>>> >>>> Regards, >>>> >>>> Vic >>> Marius -- Marius Schamschula Webmaster The Huntsville Macintosh Users Group www.hmug.org webmaster at hmug dot org marius at schamschula dot com --Apple-Mail-3--674248001 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=ISO-8859-1 Joe, Vic,

Here is what I get on my = Dual Core G5 under Mac OS X 10.4.6:

octave:1> C =3D [36 24; = 24 25];

octave:2> chol(C)

ans =3D

=A0 6=A0 4

=A0 0=A0= 3

Upper Triangular

octave:3> = version

ans =3D 2.1.72

In other words, we got to = fix this .oct problem. I just wish I knew where to = start...

On Apr 15, 2006, at 4:10 PM, Joe Koski = wrote:

Vic, Marius,

A further thought. chol = is a .oct file. Are other .oct files (trisolve.oct, etc.) working? It = may be a dynamic load library problem. The HPC version of octave was = compiled on OS X 10.3.9, and, as I recall, Vic is running 10.4.x.
=
Joe

on 4/15/06 2:53 PM, Joe Koski at jkoski11@comcast dot net wrote:
=

Vic, Marius,

= It must be the HPC version. On my home built with g95 version:

= =A0octave:1> C =3D [36 24; 24 25];
=A0octave:2> chol(C);
= =A0octave:3> chol(C)
=A0ans =3D

=A0=A0=A06 =A04
=A0=A0= =A00 =A03
=A0Upper Triangular

=A0octave:4> version
= =A0ans =3D 2.1.73
=A0

Joe

on 4/15/06 2:29 PM, = Marius Schamschula at marius173@mchsi dot com wrote:
=

Vic,

Is this = the port from hpc.sf.net, or did you compile from source?

I = have yet to get a working build of octave 2.1.73 on the Mac.

On = Apr 15, 2006, at 3:16 PM, Vic Norton wrote:

=
I had thought Octave, = version 2.1.73, was working fine on my iMac G5,
but now I'm not so = sure. Here's my problem.

I start with the matrix
=A0=A0 = octave> C =3D [36 24; 24 25];
and do
=A0=A0 octave> = chol(C);
After waiting for a while I get
=A0=A0 panic: Bus error = -- stopping myself...
=A0=A0 attempting to save variables to = `octave-core'...
=A0=A0 save to `octave-core' complete
=A0=A0 = Bus error
=A0=A0 vic$

To tell you the truth I'm not sure = what the Cholesky factor of C is
supposed to be. I do know that C =3D = R' * R where
=A0=A0 R =3D [0 3; 6 4].
I have no problem = computing another R that does this job:
=A0=A0 octave> C =3D [36 = 24; 24 25];
=A0=A0 octave> [U, S, V] =3D svd(C);
=A0=A0 = octave> R =3D sqrt(S) * V';
Now
=A0=A0 R =3D [
=A0 =A0 =A0 = =A0 =A0 -5.8067=A0 -4.6265
=A0 =A0 =A0 =A0 =A0 -1.5108 =A0 = 1.8962
=A0=A0 =A0 =A0 ]
and C =3D R' * R for this R as well. But = R =3D chol(C) simply doesn't work
for me.

Is anybody else = having any problems with "chol" or is it just my screwed
up = system?

Regards,

Vic
=

Marius

--

Marius Schamschula =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0 =A0 =A0 =A0 =A0 Webmaster

=A0 =A0 =A0 =A0 The Huntsville Macintosh Users = Group

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 www.hmug.org

webmaster at hmug dot org=A0 =A0 marius at schamschula = dot com

=

= --Apple-Mail-3--674248001-- ------------------------------------------------------------- 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 -------------------------------------------------------------