Octave 3.0.1 crashes deterministically in __qp__.cc

[Marius Schamschula]

On Jun 24, 2008, at 7:01 PM, Dmitri A. Sergatskov wrote:

> On Tue, Jun 24, 2008 at 6:51 PM, Marius Schamschula  
> <marius173 at mchsi.com> wrote:
>> John, Ben,
>> I just tried this on my (www.hmug.org) builds of octave 3.0.1;  
>> once on a G5
>> (PPC) running Mac OS X 10.4.11, and also on a MacBook Pro (Intel)  
>> running
>> 10.5.3. I get a core dump in both cases.
>
> I suspect that those were built with AltiVec and Fink's with ATLAS  
> proper...
>
> Dmitri.
> --

ATLAS proper?

Isn't the Accelerate Framework built from ATLAS sources?

The difference is that ATLAS tends to be more up to date in terms of  
bug fixes, than is the Framework.

I gave up on hand building ATLAS many years ago, i.e. build for G4  
single & dual, G5, single, dual & quad, Intel 32bit, single & dual,  
Intel 64bit dual, quad & octo...
...never mind the various sub-variants of these processors.

Who has the time, especially if Apple already has done the work in a  
more processor neutral way. Building against ATLAS also prevents  
building Octave.app (and other binary distributions, like Guarv's and  
mine), since you would need a separate build for each processor  
configuration. I currently build four versions PPC/Intel * 10.4/10.5.

BTW: note that octave with ATLAS is a Fink variant, octave can also  
be build w/o ATLAS. using the Framework.

Marius
--
Marius Schamschula
Webmaster

The Huntsville Macintosh Users Group
www.hmug.org

webmaster at hmug dot org
marius at schamschula dot com



-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://www.cae.wisc.edu/pipermail/bug-octave/attachments/20080624/6a280ff7/attachment.html