[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: octave warnings
From: |
Daniel J Sebald |
Subject: |
Re: octave warnings |
Date: |
Thu, 09 Dec 2004 02:29:11 -0600 |
User-agent: |
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3) Gecko/20041020 |
Dmitri A. Sergatskov wrote:
Daniel J Sebald wrote:
I'm having a problem with Octave 2.1.57 and 2.1.64. The former is
the binary in Fedora Core 3. The latter is what I compiled. The
behavior is the same. It could be a problem with the compiler that
is on the
Fedora Core 3 has gcc that miscompiles lapack with high optimization.
Whew. OK. Thanks. But I'm not completely following. There are many
descriptions of flags in the bugzilla reports.
You can grab lapack from the rawhide.
What is rawhide? http://www.netlib.org/lapack/lapack.tgz, Is that
rawhide? :-)
If you compile octave
w/o external lapack (and using the included one) you
have to set FFLAGS to -O1 or to -ffloat-store.
See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=138791
for details.
In bugzilla I see:
Linked with lapack rebuilt with
FFLAGS="-ffloat-store" it executes as expected.
My workaround was to recompile both lapack and octave with
FFLAGS="-O2 -ffloat-store".
Recompiling both lapack and octave with FFLAGS="-O
-ffloat-storage" seems to solve this problem.
And then I searched back through octave emails and see Quentins comment
about just getting the Fedora SRPM file and recompiling without one
having to change anything. Is that the latest and simplest solution?
I'm downloading that now and see that it is dated Nov 30th, two weeks
more recent than Quentin's posting.
Dan
- octave warnings, John W. Eaton, 2004/12/01
- Re: octave warnings, Paul Kienzle, 2004/12/02
- Re: octave warnings, Daniel J Sebald, 2004/12/08
- Re: octave warnings, Dmitri A. Sergatskov, 2004/12/09
- legend in 2.1.64, Daniel J Sebald, 2004/12/12
- Re: legend in 2.1.64, Daniel J Sebald, 2004/12/12
- Re: legend in 2.1.64, David Bateman, 2004/12/13
- feof(), ftell(), Daniel J Sebald, 2004/12/21