[gmx-users] GROMACS compilation problem on IBM AIX
Gerrit Groenhof
g.Groenhof at chem.rug.nl
Thu Feb 6 09:44:49 CET 2003
Hi again Jithesh,
> WHen you get everything running it could still be owrthwhile to
> recompile everything for power4.
ANd then compare the outcome with the outcome obtained from gromacs
compiled with -O2. That it just runs fine doesn't say anything yet. I
had the same. The simulations would just go and produce output which
did not seem too bad. But only when I compared with results obtained
on other platforms I saw that there was something wrong. The problem
with the -O3 option of the xlc compiler I experienced was that the
statements in certain functions were re-ordered. For instance in
sum_ekin (tgroup.c), it was clear that the statement were executed in
the wrong order. ONly after adding print statements, the function
would work fine with -O3. IN fact I remember that when using -O3 the
compiler by default warns the user that this can happen.
Have fun,
Gerrit
>
> > thanks & regards
> > Jithesh
> >
> > gmx-users at gromacs.org wrote:
> >
> > On Monday, Feb 3, 2003, at 09:43 US/Pacific, David van der Spoel wrote:
> >
> > > On Mon, 2003-02-03 at 15:16, Gerrit Groenhof wrote:
> > >> Hi Jithesh,
> > >>
> > >> I am not sure, but from your benchmark it seems gromacs is not
> > >> compiled properly. I had the same problem on an IBM SP3 (4?). changing
> > >> the optimization from O3 to O2 did the trick. What also works is to
> > >> change from the IBM xlc compiler to the gnu gcc compiler.
> > >>
> > > And it is also very important to check that all libraries are compiled
> > > with the same options (for hardware specifics: the processor type). If
> > > you compile part for power3 (e.g. fftw) and part for power4 (gromacs)
> > > your simulations will die.
> > >
> > > Check an option in fftw compilation, it could be that they default for
> > > the wrong architecture.,
> >
> > Yeah, but I think it will just default to pwr2 which should be fine
> > (although not as fast).
> >
> > There might indeed be a problem with the IBM compilers; if anybody can
> > provide
> > access to a power4 system I'll have a go at debugging it.
> >
> > Cheers,
> >
> > Erik
> >
> > _______________________________________________
> >
> >
> >
> > Get Your Private, Free E-mail from Indiatimes at http://email.indiatimes.com
> >
> > Buy the best in Movies at http://www.videos.indiatimes.com
> >
> > Bid for for Air Tickets @ Re.1 on Air Sahara Flights. Just log on to http://airsahara.indiatimes.com and Bid Now !
> >
> > _______________________________________________
> > gmx-users mailing list
> > gmx-users at gromacs.org
> > http://www.gromacs.org/mailman/listinfo/gmx-users
> > Please don't post (un)subscribe requests to the list. Use the
> > www interface or send it to gmx-users-request at gromacs.org.
> >
> --
> Groeten, David.
> ________________________________________________________________________
> Dr. David van der Spoel, Dept. of Cell & Mol. Biology
> Husargatan 3, Box 596, 75124 Uppsala, Sweden
> phone: 46 18 471 4205 fax: 46 18 511 755
> spoel at xray.bmc.uu.se spoel at gromacs.org http://xray.bmc.uu.se/~spoel
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
>
More information about the gromacs.org_gmx-users
mailing list