[gmx-users] why the force is so large in energy minimization

[Mark Abraham]

青 叶 wrote:
> I did a energy minimization on a system contain one [AuCl4]- and 503 
> water moleculars, by using a self-define non-bond interaction (which is 
> obtain from ab-initio calculation), my em.mdp file is as follow:
> title                  = yqq
> cpp                    = /usr/bin/cpp ; the c pre-processor
> define                 = -DFLEXILBE
> constraints            = none
> integrator             = steep
> energygrps          = AU CL OW H
> energygrp_table   = AU OW AU H CL OW CL H
> dt                     =0.002; ps!
> nsteps                 =40000
> nstlist                =10
> ns_type                =grid
> rlist                  =0.9
> coulombtype            =User
> rcoulomb               =0.9
> vdwtype                =User
> rvdw                   =1.0
> fourierspacing         =0.12
> fourier_nx             =0
> fourier_ny             =0
> fourier_nz             =0
> pme_order              =4
> ewald_rtol             =1e-5
> optimize_fft           =yes
> ;
> ;                     Energy minimizing stuff
> ;
> emtol                  =500.0
> emstep                 =0.005
>  
> and meanwhile, I have modifide the ffG43a1.rtp, ffG43a1bon.itp, 
> ffG43a1nb.itp files in order to include the relative parameters. Then 
> add a table.xvg (actually the default 6-12 Lennard Jones potential form, 
> I just renamed it so it will be read for describing the interaction of 
> atom pairs which are not list in the  energygrp_table section) and four 
> table****.xvg contain the interaction information of AU OW, AU CL, CL 
> OW, CL H.
> but unfortunatly, the corresponding em.log file is follow:
>  
>  
> Stepsize too small, or no change in energy.
> Converged to machine precision,
> but not to the requested precision Fmax < 500
> 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)
> Steepest Descents converged to machine precision in 12 steps,
> but did not reach the requested Fmax < 500.
> Potential Energy  = -3.5686449e+04
> Maximum force     =  1.3609032e+04 on atom 1236
> Norm of force     =  1.6019813e+03
>  M E G A - F L O P S   A C C O U N T I N G
>    RF=Reaction-Field  FE=Free Energy  SCFE=Soft-Core/Free Energy
>    T=Tabulated        W3=SPC/TIP3p    W4=TIP4p (single or pairs)
>    NF=No Forces
>  Computing:                         M-Number         M-Flops  % Flops
> -----------------------------------------------------------------------
>  VdW(T)                             1.011752          54.635    16.6
>  Coul(T) + VdW(T)                   2.659114         180.820    55.0
>  Outer nonbonded loop               0.261041           2.610     0.8
>  NS-Pairs                           4.096803          86.033    26.2
>  Reset In Box                       0.012096           0.036     0.0
>  Shift-X                            0.018168           0.109     0.0
>  CG-CoM                             0.018168           0.055     0.0
>  Bonds                              0.000048           0.003     0.0
>  Angles                             0.000072           0.012     0.0
>  Impropers                          0.000024           0.005     0.0
>  Virial                             0.018708           0.337     0.1
>  Settle                             0.012072           3.899     1.2
> -----------------------------------------------------------------------
>  Total                                               328.553   100.0
> -----------------------------------------------------------------------
> 
>      R E A L   C Y C L E   A N D   T I M E   A C C O U N T I N G
>  Computing:         Nodes     Number     G-Cycles    Seconds     %
> -----------------------------------------------------------------------
>  Neighbor search        1         12        0.844        0.8    77.7
>  Force                  1         12        0.180        0.2    16.5
>  Constraints            1         23        0.004        0.0     0.4
>  Rest                   1                   0.059        0.1     5.4
> -----------------------------------------------------------------------
>  Total                  1                   1.086        1.0   100.0
> -----------------------------------------------------------------------
>                NODE (s)   Real (s)      (%)
>        Time:      1.000      1.000    100.0
>                (Mnbf/s)   (MFlops)   (steps/hour)
> Performance:      3.671    328.553        43200.0
> Finished mdrun on node 0 Wed Oct 21 23:38:48 2009
>  
> the force is very large and I have tried to adjust "emstep" and other 
> options but seems has no influence on the result, can anyone help me on 
> this problem, any suggestion will be very, very grateful. Thank you.
> By the way, I have ajusted the No. of water moleculars in the simulation 
> box but almost has no infulence on the result.

The magnitude of the PE is such that you've probably not made any 
catastrophic errors. Make sure you get the sign of the table columns 
correct for various derivatives correct. Otherwise, have a look at atom 
1236 and see what might be wrong in that region.

Mark