Hello Satish and Matthias,
Here's what I've tried so far with PETSc 2.3.2:
* -1 Patch level 6: HPPA segfaulted with python2.3, took over 150
minutes with python2.4.
* -2 Patch level 8, python 2.4: ARM died during checkDynamicLinker
test; HPPA still took over 150 minutes (m68k dies while
compiling).
* -3 Patch level 8, python 2.4, --useThreads 0: HPPA died at start
of config: "Fatal Python error: PyEval_RestoreThread: NULL
tstate"; ARM outcome unknown.
Build logs: http://buildd.debian.org/build.php?arch=&pkg=petsc
-Adam
On Thu, 2007-01-11 at 09:54 -0600, Satish Balay wrote:
> Adam,
>
> Can you try the latest patchlevel and see if it works?
>
> There was a change in the threads code on the petsc side - for a
> performance problem with threads. [but 150min is a bit too long..]
>
> Satish
>
> On Thu, 11 Jan 2007, Adam C Powell IV wrote:
>
> > Yes, it is reproducible with python2.4, as recently as four days ago:
> > http://buildd.debian.org/fetch.cgi?&pkg=petsc&ver=2.3.2-1&arch=hppa&stamp=1168166295&file=log
> >
> > ...
> > Checking for already installed source dependencies...
> > debhelper: missing
> > python: already installed (2.4.4-2 >= 2.2 is satisfied)
> > ...
> > PETSC_DIR=/build/buildd/petsc-2.3.2 ./config/configure.py \
> > --with-cc=gcc --with-fc=g77 --with-debugging=1 \
> > --with-mpi-dir=/usr/lib/mpich \
> > --with-blas-lib=-lblas-3 --with-lapack-lib=-llapack-3
> > make: *** wait: No child processes. Stop.
> > make: *** Waiting for unfinished jobs....
> > make: *** wait: No child processes. Stop.
> > Build killed with signal 15 after 150 minutes of inactivity
> >
> > This works fine on all of the other arches, even m68k and ARM which I
> > believe are significantly slower (though I'm not sure why neither m68k
> > nor ARM tried 2.3.2-1). This leads me to believe that it's a python bug
> > on HPPA. What makes you think it's not?
> >
> > -Adam
> >
> > On Thu, 2007-01-11 at 10:58 +0100, Matthias Klose wrote:
> > > reassign 354139 petsc
> > > thanks
> > >
> > > reassigning to petsc; please check if that is reproducible with
> > > python2.4 (if it's a python bug at all).
--
GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6
Welcome to the best software in the world today cafe!
http://www.take6.com/albums/greatesthits.html
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
