[gmx-users] different results in different machines

[XAvier Periole]

On Feb 11, 2010, at 7:40 PM, Rebeca García Fandiño wrote:

> Hi,
> I am doing PMF calculations using Gromacs 3.3.3, and I have found  
> very different results using 2 different machines.  Attached you can  
> see that the graphics are really different for both cases (about 10  
> KJ/mol!!). The version of Gromacs (3.3.3) and number of processors  
> used (2) is the same for both calculations. The .tpr files and all  
> the starting files are identical for both calculations. The only  
> thing that changes is the machine where they were carried out.
> I have checked the energy files for both cases, and they are  
> different, but taking into account the stochastic effect, I suppose  
> it is normal. There are slight differences in the energies in the  
> step 0 for some of the windows (but in the 5th decimal number, so  
> they should not be important, I suppose).
Well this might be a problem but the 5th decimal is pretty far.
> Any idea of the cause of the problem? Has anyone experienced similar  
> problems? Any suggestion will be much appreciated.
> Thanks a lot for your help, in advance.
the only thing you have left is that you have not sampled enought ... so
you have not reached convergence.
Did you try to repeat the calculation on the ame machine changing the
initial velocities ?

> Cheers,
>
> Rebeca Garcia
>
>
> Elige un juego de Messenger. Reta a un amigo. ¡Pruébalo ya!  
> <comparation_neuron_ngs.pdf>--
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before  
> posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100211/bd79cb0f/attachment.html>