Re: [ESPResSo-users] Fwd: question on restarting after crash

[Axel Arnold]

Hi Marcello,

if the forces are NaN and not zero, they get recomputed, but one of the 
interactions you use is not computing correctly. Can you provide a 
minimal testcase? Also, you can just switch off interactions one after 
the other to see which one actually computes the NaNs.

Axel

On 29.04.14 09:14, Marcello Sega wrote:
> Dear all,
>
> I discovered I was not subscribed to the user mailing list, so most
> probably my message below was trashed automatically. I resend it now,
> my apologies if somebody is getting it twice:
>
> I'm currently lost with the problem below, maybe some of you have an
> idea of what is going on....
>
> In a cycle like this:
>
> while { 1 } {
>   kill_particle_motion
>   kill_particle_forces
>   save_positions
>    if { [catch { integrate 20 } res ] } {
>        invalidate_system
>        kill_particle_motion
>        kill_particle_forces
>        load_positions
>        integrate 0
>    }
> }
>
> the [integrate 0] reports that forces are NaN (it's not the first
> iteration in the while loop, the [ integrate 20 ] is successful for
> some time before the crash)
>
> Am I missing something? when I do [integrate 0], all positions,
> velocities and forces should (I think) be the same as when the
> integration was successful for the last time...
>
> Any advice is greatly appreciated...
>
> Kind Regards,
>
> Marcello
>
>
>
>
>
>
>
>
> --
> Institut für Computergestützte Biologische Chemie
> University of Vienna
> PGP public key on MIT server http://goo.gl/zlIZix


--
JP Dr. Axel Arnold
ICP, Universität Stuttgart
Allmandring 3
70569 Stuttgart, Germany
Email: address@hidden
Tel: +49 711 685 67609