[gmx-users] The simulation of big protein in implicit solvent

[Justin Lemkul]

On 9/1/16 5:49 AM, Yi Isaac Yang wrote:
> Dear all,
>
> Recently, I want to simulate a big protein (about 90 residues) in GBSA implicit
> solvent using gromacs 5.1.2. However, when I want to perform the minimization,
> the program shows that:
>
> _Fatal error:__
> __Domain decomposition does not support simple neighbor searching, use grid
> searching or run with one MPI rank__
> __For more information and tips for troubleshooting, please check the GROMACS__
> __website at http://www.gromacs.org/Documentation/Errors_
>
> And the strange thing is, I can use the mdp file to minimization a small protein
> (6 residues, using the same force field), every thing is OK.
>
> If I delete "ns-type=simple" in mdp file, the minimization of  small protein is
> still OK. And the minimization of big protein is still not work:
> _
> __Fatal error:__
> __Domain decomposition does not work with nstlist=0__
> __For more information and tips for troubleshooting, please check the GROMACS__
> __website at http://www.gromacs.org/Documentation/Errors_
>
> As no-cutoff is needed in this simulation, I can't delete "_nstlist=0_".
>
> So you know how to solve this problem?
>

Basically, you can't parallelize via DD.  See the other thread to which I just 
posted a reply.  Implicit simulations don't scale well and you are therefore 
limited in the types of algorithms that are supported.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================