[gmx-users] minimzation problem with a heterogeneous system

[Mark Abraham]

Payman Pirzadeh wrote:
> Hello,
> I have a configuration in which half of the box is liquid water and the
> other half is ice. This configuration is the output of another simulation in
> which the mentioned config worked (and still works fine). But when I start
> minimizing it (I started minimization since my first round of MD run crashed
> due to messages indicative of clashes). But minimation ends with the
> following message:
> 
> Steepest Descents:
>    Tolerance (Fmax)   =  1.00000e-06
>    Number of steps    =       500000
> Step=    1, Dmax= 1.0e-01 nm, Epot=  6.09464e+11 Fmax= 1.31718e+14, atom=
> 2938
> Step=    2, Dmax= 1.2e-01 nm, Epot=  1.15027e+09 Fmax= 2.00242e+11, atom=
> 2576
> Step=    3, Dmax= 1.4e-01 nm, Epot=  1.56529e+08 Fmax= 2.20702e+09, atom=
> 11062
> 
> t = 0.004 ps: Water molecule starting at atom 16078 can not be settled.
> Check for bad contacts and/or reduce the timestep.
> Wrote pdb files with previous and current coordinates
> Step=    4, Dmax= 1.7e-01 nm, Epot=  2.11002e+10 Fmax= 5.75671e+13, atom=
> 14986
> t = 0.005 ps: Water molecule starting at atom 14902 can not be settled.
> Check for bad contacts and/or reduce the timestep.
> Wrote pdb files with previous and current coordinates
> Step=    5, Dmax= 8.6e-02 nm, Epot=  1.22915e+07 Fmax= 4.57843e+07, atom=
> 6280
> 
> t = 0.006 ps: Water molecule starting at atom 15952 can not be settled.
> Check for bad contacts and/or reduce the timestep.
> Wrote pdb files with previous and current coordinates
> Step=    7, Dmax= 5.2e-02 nm, Epot=  5.50887e+06 Fmax= 1.19611e+08, atom=
> 7348
> Step=    8, Dmax= 6.2e-02 nm, Epot=  4.29974e+06 Fmax= 1.39548e+07, atom=
> 12066
> Step=    9, Dmax= 7.5e-02 nm, Epot=  1.93270e+06 Fmax= 2.81542e+07, atom=
> 7204
> 
> t = 0.010 ps: Water molecule starting at atom 7300 can not be settled.
> Check for bad contacts and/or reduce the timestep.
> Wrote pdb files with previous and current coordinates
> Step=   26, Dmax= 1.4e-06 nm, Epot=  2.53306e+06 Fmax= 8.74996e+06, atom=
> 15953
> Stepsize too small, or no change in energy.
> Converged to machine precision,
> but not to the requested precision Fmax < 1e-06
> 
> Double precision normally gives you higher accuracy.
> You might need to increase your constraint accuracy, or turn
> off constraints alltogether (set constraints = none in mdp file)
> 
> writing lowest energy coordinates.
> 
> Steepest Descents converged to machine precision in 27 steps,
> but did not reach the requested Fmax < 1e-06.
> Potential Energy  =  1.9327006e+06
> Maximum force     =  2.8154184e+07 on atom 7204
> Norm of force     =  4.1820662e+05
> 
> I checked the output .pdb files, and they show box is cracking from the
> position of the ice-water interface! Also, the ice part looks like a
> complete mess! (like it's not going to be ice anymore). I checked the .itp
> file to make sure are the variables are correct. Nothing looked wrong to me.
> I checked the mailing list about minimization. Just remember this water
> model is not flexible. 

All we can say is the something is broken with your model or initial 
configuration. Simplify it into small, homogeneous systems until you can 
diagnose the source of the problem.

Mark