[gmx-users] Segmentation fault (core dumped)

[Justin Lemkul]

On 6/6/13 4:45 AM, Ishwor Poudyal wrote:
>
> 		Dear all
>
> I want to study the diffusion coeffcient of CO in water.  I have done the energy minimization step and got the problem segmentation fault during Equilibration. I am confused whether my input file has
> got some error or the error is in the processing machine.I have made the
>   following .mdp file
>
> ;
>
> ;PREPROCESSING parameters
>
> cpp                 =  /lib/cpp
>
> define              =  -DFLEX_SPCE
>

Here's the first suspect.  The water models in Gromacs were intended to be 
rigid.  Flexibility should only be used during EM, and only if necessary to 
improve the outcome.  Running MD with flexible water is not advised.

> integrator          =  md
>
> dt		    =.002
>
> nsteps              = 25000000
>
> nstcomm		    = 1
>
>
> ;OUPUT CONTROL parameters.
>
> nstxout             =  250
>
> nstvout             =  1000
>
> nstfout             =  0
>
> nstlog              =  100
>
> nstenergy           =  100
>
> energygrps	    =  system
>
> ;NEIGHBOUR SEARCHING parameters.
>
> nstlist             =  10
>
> ns_type             =  grid
>
> rlist               =  1.0
>
> ;ELECTROSTATIC and VdW parameters.
>
> rcoulomb            =  1.0
>
> rvdw                =  1.0
>
> epsilon-r           =  1
>
> ;BERENDSEN TEMPERATURE COUPLING is on in two groups
>
> Tcoupl              =  berendsen
>
> tc-grps		    =  system
>
> tau_t               =  0.01	
>

This is a very restrictive value of tau_t.  Normally something like 0.1 or 0.5 
is more appropriate.

> ref_t               =  300	
>
> ;PRESSURE COUPLING is on
>
> Pcoupl              =  berendsen
>
> tau_p               =  0.1
>

Again, very restrictive, especially for pressure.  Try 1.0 or 2.0 instead.

> compressibility     =  4.6e-5
>
> ref_p               =  1.0
>
> ;SIMULATED ANNEALING parameters are not specified.
>
> ;GENERATE VELOCITIES is on at 300 K.
>
> gen_vel             =  yes;     ; generate initially
>
> gen_temp            =  300
>
> gen_seed            =  173259   ;give different values for different trials.
>
> ;BONDS parameters
>
> pbc		    = xyz		; 3-D PBC
>
> constraints         = all-bonds
>
> constraint-algorithm = shake
>
> unconstrained-start  = no
>
>   The box size is 2.1 nm.
>

You're playing with fire here - if the box deviates just a little bit due to 
pressure oscillations, your 1.0 nm will begin to double-count interactions and 
violate the minimum image convention.  In that case, your trajectory is junk.

-Justin

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

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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