[gmx-users] Who uses Gromacs on Power4 ?

[Campbell Millar]

Hi Folks,

I haven't tried GROMACS on power 4 as yet but I have had similar 
problems with the IBM xlc compilers in the past. Have you tried the 
-qstrict option ? Some of the more aggressive optimisation have been 
known to completely destroy the floating  point accuracy on the power3 
at least. I would try building with -O2 and see if that makes a 
difference.

The only other thing I can think of is that I once had a similar 
problem because an old version of libc was hanging around and realloc 
got broken.

Hope this is of some small help,

Campbell

On 4 Jun 2004, at 23:57, Marc Baaden wrote:

>
> Hi again,
>
> thanks for all your answers and suggestions.
>
> I have recompiled with the suggested options, but with continuing 
> (although
> for now different) error.
>
> I wanted to choose an extremely simple test system, and I used the 
> water
> example from the tutorial, trying to run on just a single processor.\
> The error is different, but again occurs before
> or at the very first time step:
>
> [..]
> starting mdrun 'Pure Water'
> 10000 steps,     20.0 ps.
>
> step 0
> Warning: Only triclinic boxes with the first vector parallel to the 
> x-axis and the second vector in the xy-plane are supported.
>          Box (3x3):
>             Box[    0]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    1]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    2]={        NaNQ,         NaNQ,         NaNQ}
>          Can not fix pbc.
> Warning: Only triclinic boxes with the first vector parallel to the 
> x-axis and the second vector in the xy-plane are supported.
>          Box (3x3):
>             Box[    0]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    1]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    2]={        NaNQ,         NaNQ,         NaNQ}
>          Can not fix pbc.
> Warning: Only triclinic boxes with the first vector parallel to the 
> x-axis and the second vector in the xy-plane are supported.
>          Box (3x3):
>             Box[    0]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    1]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    2]={        NaNQ,         NaNQ,         NaNQ}
>          Can not fix pbc.
> Warning: Only triclinic boxes with the first vector parallel to the 
> x-axis and the second vector in the xy-plane are supported.
>          Box (3x3):
>             Box[    0]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    1]={        NaNQ,         NaNQ,         NaNQ}
>             Box[    2]={        NaNQ,         NaNQ,         NaNQ}
>          Can not fix pbc.
>
>
> So I think the tpr file is read incorrectly (the very same file runs 
> absolutely
> fine with eg mdrun from gmx321 under linux).
>
> Could this have something todo with 32 vs 64 bit ? (eg 
> mis-interpreting/reading
> of the tpr data)
>
> I have tried both a tpr generated on a linux box and a tpr generated 
> on the
> power4 .. same error.
>
> Cheers,
>   Marc
>
>
> -- 
>  Dr. Marc Baaden  - Institut de Biologie Physico-Chimique, Paris
>  mailto:baaden at smplinux.de      -      http://www.marc-baaden.de
>  FAX: +49 697912 39550  -  Tel: +33 15841 5176 ou +33 609 843217
>
>
> _______________________________________________
> 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.
>
__________________________
Dr Campbell Millar,
University of Glasgow,
Device Modelling Group,
Oakfield Avenue,
Glasgow G12 8LT
tel: +44 141 330 4792  fax: +44 141 330 4907

"Diplomacy is the art of saying 'nice doggy', till you can find a rock"
Larry Niven.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2375 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20040608/7e7685f8/attachment.p7s>