compilation on a power5

[Axel]

On Apr 7, 2:03 pm, Marius Retegan <marius.s... at gmail.com> wrote:
> So this means that the compiler might be faulty?

not sure if this is a big help, but there is a v0.91 g95 available for
powerpc-aix. perhaps that might work as a control sample.
it seems very likely that the xlf compiler is trying very hard to
optimize
something that doesn't need optimizing and is not respecting the
commandline flags (or rather, there is some other, hard to find flag
that does the job).

cheers,
   axel.

> Marius
>
> On Tue, Apr 7, 2009 at 7:33 PM, Matt W <MattWa... at gmail.com> wrote:
>
> > I take back previous comments - having updated the CVS I get an
> > identical error to you.
>
> > Matt
>
> > On Apr 7, 1:21 pm, Marius Retegan <marius.s... at gmail.com> wrote:
> > > Hi Matt,
>
> > > On Tue, Apr 7, 2009 at 1:39 PM, Matt W <MattWa... at gmail.com> wrote:
>
> > > > Strange.  Can you compile the md_run.f90 manually (or copy the
> > > > unoptimised object/module files) and continue the compilation or just
> > > > continue the compilation?  With the arch file you've got the md_run
> > > > should be being compiled without optimisation anyway, so perhaps the
> > > > real error is elsewhere.
>
> > > It is indeed strange. I've tried compiling the md_run.f90 by itself. Same
> > > error. I don't understand why is he complaining about memory, when it's
> > > clear to me that it has enough? I tried increasing the verbosity of the
> > > output, but nothing new appears.
>
> > > > I've been able to compile OK (most of the time) - often the
> > > > compilation times out in the file you refer to, but you seem to have
> > > > more privileges than me.
>
> > > > uclmbw:uclmbw> xlf -qversion
> > > > IBM XL Fortran Enterprise Edition V10.1 for AIX
> > > > Version: 10.01.0000.0009
>
> > > Here I get Version: 10.01.0000.0000
>
> > > > uclmbw:uclmbw> uname -a
> > > > AIX l1f401 3 5 00CD7EAF4C00
> > > > uclmbw:uclmbw> oslevel
> > > > 5.3.0.0
>
> > > Would it be possible to send me the executable? Again, I'm not sure that
> > is
> > > going to work, but at least I can try
>
> > > > Matt
>
> > > Thanks,
> > > Marius