[gmx-users] increasing forces during energy minimization

[Ingo Appenzeller]

Dear GROMACS community,

I’ve got some problems getting my energy minimization with mdrun work right for my solution barium-sulfate-ions in water (tip4p). After some steps of calculation (integrator=steep) an decreasing forces, the forces begin to fastly increase again by multiple dimensions. The calculation ist stopped with Epot = –nan. Here the complete output after starting grompp and mdrun: 

“Back Off! I just backed up md.log to ./#md.log.11#

Reading file topol.tpr, VERSION 4.6.5 (single precision)

Using 1 MPI thread

Using 2 OpenMP threads



Back Off! I just backed up traj.trr to ./#traj.trr.10#



Back Off! I just backed up ener.edr to ./#ener.edr.10#



Steepest Descents:

   Tolerance (Fmax)   =  1.00000e+01

   Number of steps    =        50000

Step=    0, Dmax= 1.0e-03 nm, Epot= -1.32579e+05 Fmax= 1.05935e+04, atom= 5893

Step=    1, Dmax= 1.0e-03 nm, Epot= -1.33511e+05 Fmax= 9.70591e+03, atom= 5893

Step=    2, Dmax= 1.2e-03 nm, Epot= -1.34590e+05 Fmax= 8.73726e+03, atom= 5893

Step=    3, Dmax= 1.4e-03 nm, Epot= -1.35838e+05 Fmax= 7.86312e+03, atom= 12489

Step=    4, Dmax= 1.7e-03 nm, Epot= -1.37251e+05 Fmax= 6.97747e+03, atom= 12489

Step=    5, Dmax= 2.1e-03 nm, Epot= -1.38845e+05 Fmax= 6.07222e+03, atom= 13045

Step=    6, Dmax= 2.5e-03 nm, Epot= -1.40644e+05 Fmax= 5.17057e+03, atom= 13045

Step=    7, Dmax= 3.0e-03 nm, Epot= -1.42687e+05 Fmax= 4.29438e+03, atom= 13045

Step=    8, Dmax= 3.6e-03 nm, Epot= -1.45028e+05 Fmax= 3.46452e+03, atom= 13045

Step=    9, Dmax= 4.3e-03 nm, Epot= -1.47752e+05 Fmax= 2.69784e+03, atom= 13045

Step=   10, Dmax= 5.2e-03 nm, Epot= -1.50986e+05 Fmax= 2.12213e+03, atom= 6097

Step=   11, Dmax= 6.2e-03 nm, Epot= -1.54713e+05 Fmax= 1.67687e+03, atom= 6097

Step=   12, Dmax= 7.4e-03 nm, Epot= -1.58896e+05 Fmax= 1.29268e+03, atom= 6097

Step=   13, Dmax= 8.9e-03 nm, Epot= -1.63532e+05 Fmax= 9.75267e+02, atom= 10463

Step=   14, Dmax= 1.1e-02 nm, Epot= -1.68496e+05 Fmax= 9.16984e+02, atom= 11

Step=   15, Dmax= 1.3e-02 nm, Epot= -1.72701e+05 Fmax= 2.03812e+03, atom= 11

Step=   16, Dmax= 1.5e-02 nm, Epot= -1.74483e+05 Fmax= 5.56818e+03, atom= 11

Step=   17, Dmax= 1.8e-02 nm, Epot= -1.75619e+05 Fmax= 2.56424e+04, atom= 15

Step=   18, Dmax= 2.2e-02 nm, Epot= -1.90325e+05 Fmax= 5.23639e+06, atom= 11

Step=   24, Dmax= 8.3e-04 nm, Epot= -2.09743e+05 Fmax= 2.60194e+07, atom= 11

Step=   25, Dmax= 1.0e-03 nm, Epot= -2.48560e+05 Fmax= 1.15010e+08, atom= 15

Step=   33, Dmax= 9.4e-06 nm, Epot= -2.50790e+05 Fmax= 1.22044e+08, atom= 15

Step=   36, Dmax= 2.8e-06 nm, Epot= -2.51480e+05 Fmax= 1.24264e+08, atom= 15

Step=   38, Dmax= 1.7e-06 nm, Epot= -2.51892e+05 Fmax= 1.25597e+08, atom= 15

Step=   40, Dmax= 1.0e-06 nm, Epot=         -nan Fmax= 3.38430e+03, atom= 16

Energy minimization has stopped, but the forces have not converged to the

requested precision Fmax < 10 (which may not be possible for your system). It

stopped because the algorithm tried to make a new step whose size was too

small, or there was no change in the energy since last step. Either way, we

regard the minimization as converged to within the available machine

precision, given your starting configuration and EM parameters.



Double precision normally gives you higher accuracy, but this is often not

needed for preparing to run molecular dynamics.

You might need to increase your constraint accuracy, or turn

off constraints altogether (set constraints = none in mdp file)



writing lowest energy coordinates.



Back Off! I just backed up outSULWA2min.gro to ./#outSULWA2min.gro.3#



Steepest Descents converged to machine precision in 41 steps,

but did not reach the requested Fmax < 10.

Potential Energy  = -2.5189184e+05

Maximum force     =  1.2559665e+08 on atom 15

Norm of force     =  1.3859668e+06



NOTE: 14 % of the run time was spent in pair search,

      you might want to increase nstlist (this has no effect on accuracy)“





Mdrun seems to skip some calculation steps after the 18th step as well. 

My commands were as follows:



grompp           -f minim.mdp

                        -c outBASULWA.gro

                        -p topolSUL1.top

                        -o topolSULWAmin.tpr



mdrun             -v

                        -c outBASULWAmin.gro

                        -s topolSULWAmin.tpr



I’d be quite thankful for any hint.



With kind regards



Ingo Appenzeller

Karlsruhe Institute of Technology