[gmx-users] segementation fault- water molecule can not be settled- free mdrun error (repica exchange)

Justin Lemkul jalemkul at vt.edu
Mon Jan 13 15:16:30 CET 2014

On 1/13/14, 8:44 AM, Guo, Yanan wrote:
> Dear Justin,
> Thanks a lot for your help.
> According to your suggestion, I changed my pressure coupling settings:
> ; Pressure coupling
> pcoupl          = no
> pcoupltype      = isotropic
> tau_p           = 2.0
> ref_p           =
> compressibility =
> Then there is no “segementation fault”.Thank you very much!  (I also made a test by reducing the time step, but it didn't work.)
> But I have a question here. Given the whole protocol of my simulation:
> 1.  energy minimization of the system
> 2. position restrained equilibration of the system in NVT ensemble and NPT ensemble.
> 3. free MD simulation (replica exchange and simulated annealing)
> To avoid the collapse of the system with a sudden relax, I released the position restraints step by step in NPT ensemble before free MD simulation.  So I am wondering is it proper to turn off the pressure coupling (pcoupl = 0) in a free MD simulation? Especially in the case that I want to treat this free MD as the production MD simulation? Or we cann't use replica exchange as a production MD simulation?

It depends on what you hope to model.  If you want to simulate under an NPT 
ensemble, you need to find a way to make that work (various suggestions have 
been posted over the years and are in the list archive).  Whatever your settings 
were during NPT REMD were not suitable stable.  It can be very challenging to 
use an NPT ensemble during REMD for reasons stated previously.  You should not 
blindly settle for something that "just works" if it does not represent the 
physical reality you wish to model.



Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441


More information about the gromacs.org_gmx-users mailing list