[gmx-users] Implicit solvent in Gromacs 5.0.3

[Justin Lemkul]

On 1/27/15 7:29 PM, Stephen Fried wrote:
> Hello all --
>
> I recently installed a distribution of the new GROMACS 5, and I am finding
> myself in somewhat of a rut to perform implicit solvent simulations on the
> mpi-build.
>
> To put it simply, if I specify the cutoff-scheme as Verlet, GROMACS fails
> at grompp, because "Implicit solvent is not (yet) supported with the with
> Verlet lists."
>
> However, if I change cutoff-scheme to group, then I run into trouble upon
> mdrun, because "OpenMP threads have been requested with cut-off scheme
> Group, but these are only supported with cut-off scheme Verlet."
>
> Grr!  Am I missing something, or is the simulation I want to do currently
> not possible?  And if that's the case, would you recommend returning to
> GROMACS 4, or what else?
>

If you built with MPI, you can use mdrun -ntmpi instead of allowing -nt to 
simply allocate as it thinks best.  I doubt you'll get much of a boost in 
performance via parallelization with implicit solvent, anyway.

But in truth, there is little developer support for implicit solvent and there 
are a lot of changes within the code that are always ongoing (as you're 
noticing), especially as the group scheme is phased out entirely.

-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

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