Did you change your time step when you turned off shake? I would expect
integration errors (resulting in this SHAKE error) if your time step was too
large for some of your higher-frequency motions.
Also, repeating what Dan said, you should *always* minimize your starting
structure so that large initial forces don't blow your system apart. Also,
do your parallel pmemd tests pass? If so, I would suggest refining your
starting structure and re-running your dynamics.
HTH,
Jason
On Thu, Aug 18, 2011 at 1:01 AM, Danny Xu <quantum_mania.yahoo.com> wrote:
> Dan,
>
> pmemd.MPI gave me SHAKE error at the very start, the parallel job crashed
> immediately. Serial pmemd did not have this problem at all. It's running,
> I'm pretty sure it will finish in the end.
>
> If the structure was not sufficiently minimized, I don't understand why
> serial pmemd worked just fine with and without SHAKE... It feels like there
> may be some bug in pmemd.MPI.
>
>
> DX
>
>
>
> ________________________________
> From: Daniel Roe <daniel.r.roe.gmail.com>
> To: Danny Xu <quantum_mania.yahoo.com>; AMBER Mailing List <
> amber.ambermd.org>
> Sent: Wednesday, August 17, 2011 10:20 PM
> Subject: Re: [AMBER] Coordinate resetting cannot be accomplished, deviation
> is too large
>
> Quick question - when you say serial pmemd works fine, do you mean it
> completes all 850000 steps? Is it able to complete more steps than the
> parallel run? Usually when you receive a message like this:
>
> On Wed, Aug 17, 2011 at 10:13 PM, Danny Xu <quantum_mania.yahoo.com>
> wrote:
> > Coordinate resetting cannot be accomplished,
> > deviation is too large
>
> and see behavior like this:
>
> > If I disable SHAKE (ntc=1, ntf=1), pmemd.MPI job will run, but energies
> blow up in a short period of time.
>
> it means your starting coordinates are not well-minimized.
>
> -Dan
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
> _______________________________________________
> 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 Thu Aug 18 2011 - 08:00:03 PDT
