I'd say this is fairly inconsequential. This doesn't happen very often
because 200M steps is incredibly large to run in a single simulation.
Furthermore, 0.002 doesn't exist in 64-bit binary (or any finite-bit
binary), so after enough time, adding 0.002 over and over and over will
inevitably lead to this type of artifact (though I admit, even 200 million
steps is quicker than I would have anticipated the 4th decimal place being
impacted, but not by too much).
This could almost certainly be remedied by breaking your simulation up into
multiple, smaller steps, since the "error" accumulation (caused by
continually adding up the residual difference between the "real" 0.002 and
as close as a computer can get to it) will be reset along with the
simulation.
As others have suggested, this is not something to worry about, nor is
there much we can do short of defining time (and dt) with a higher level of
precision (like real*16 or something). That brings its own problems,
though.
HTH,
Jason
On Wed, Jan 4, 2012 at 3:04 AM, Dmitry Mukha <dvmukha.gmail.com> wrote:
> Error is probably being accumulated. I found another event in this run, but
> it doesn't concern a large jump.
> .
> NSTEP =214629000 TIME(PS) = 429357.999 TEMP(K) = 300.23 PRESS =
> 0.0
> Etot = -43915.5075 EKtot = 10320.2637 EPtot =
> -54235.7712
> BOND = 118.7568 ANGLE = 317.6830 DIHED =
> 414.5547
> 1-4 NB = 117.3973 1-4 EEL = 940.9223 VDWAALS =
> 7502.6621
> EELEC = -63647.7475 EHBOND = 0.0000 RESTRAINT =
> 0.0000
>
> ------------------------------------------------------------------------------
>
> wrapping first mol.: 223.28419467423717 -28.706496795098609
> 49.721126343645942
>
> NSTEP =214629500 TIME(PS) = 429358.998 TEMP(K) = 301.04 PRESS =
> 0.0
> Etot = -43910.8599 EKtot = 10348.0127 EPtot =
> -54258.8726
> BOND = 119.5408 ANGLE = 320.1453 DIHED =
> 410.9057
> 1-4 NB = 126.3671 1-4 EEL = 930.5164 VDWAALS =
> 7540.1994
> EELEC = -63706.5472 EHBOND = 0.0000 RESTRAINT =
> 0.0000
>
> ------------------------------------------------------------------------------
>
>
>
> 2012/1/3 Ross Walker <ross.rosswalker.co.uk>
>
> >
> > If the issue seems to be more major than this - I.e. the time actually
> > 'jumped' several ps then let me know.
> >
> > All the best
> > Ross
> >
> > /\
> > \/
> > |\oss Walker
> >
> > ---------------------------------------------------------
> > | Assistant Research Professor |
> > | San Diego Supercomputer Center |
> > | Adjunct Assistant Professor |
> > | Dept. of Chemistry and Biochemistry |
> > | University of California San Diego |
> > | NVIDIA Fellow |
> > | http://www.rosswalker.co.uk | http://www.wmd-lab.org/ |
> > | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> > ---------------------------------------------------------
> >
> > Note: Electronic Mail is not secure, has no guarantee of delivery, may
> not
> > be read every day, and should not be used for urgent or sensitive issues.
> >
> >
> >
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
>
>
>
> --
> Sincerely,
> Dmitry Mukha
> Institute of Bioorganic Chemistry, NAS, Minsk, Belarus
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jan 04 2012 - 08:00:03 PST
