[gmx-users] Minimization error

[Ahmed Mashaly]

Hello,
What should be the emtol value for a big system (+1600 residue, +50K water)? the default is 10 


I tried emtol = 10 with cg and had this error in the beginnings
step -1: One or more water molecules can not be settled.Check for bad contacts and/or reduce the timestep if appropriate.Wrote pdb files with previous and current coordinates
Fatal error:
The coordinates could not be constrained. Minimizer 'cg' can not handleconstraint failures, use minimizer 'steep' before using 'cg'.


I tried etmol = 10 with steep and got this message after +9000 step, with em.gro produced
Energy minimization has stopped, but the forces have not converged to therequested precision Fmax < 10 (which may not be possible for your system). Itstopped because the algorithm tried to make a new step whose size was toosmall, or there was no change in the energy since last step. Either way, weregard the minimization as converged to within the available machineprecision, given your starting configuration and EM parameters.
Double precision normally gives you higher accuracy, but this is often notneeded for preparing to run molecular dynamics.You might need to increase your constraint accuracy, or turnoff constraints altogether (set constraints = none in mdp file)
writing lowest energy coordinates.
Steepest Descents converged to machine precision in 9400 steps,but did not reach the requested Fmax < 10.Potential Energy  = -3.8866718e+06Maximum force     =  1.2978005e+03 on atom 21160Norm of force     =  4.9661236e+00
Simulation ended prematurely, no performance report will be written.


And when I tried cg with emtol = 10 using the last em.gro, I had this message after few tens of steps
step -1: One or more water molecules can not be settled.Check for bad contacts and/or reduce the timestep if appropriate.Wrote pdb files with previous and current coordinates
Fatal error:The coordinates could not be constrained. Minimizer 'cg' can not handleconstraint failures, use minimizer 'steep' before using 'cg'.


but when I tried cg with emtol=100, I get it done, with .gro file
Step 0, Epot=-3.885809e+06, Fnorm=5.670e+01, Fmax=6.994e+02 (atom 21160)Step 1, Epot=-3.886087e+06, Fnorm=3.594e+01, Fmax=3.332e+02 (atom 7929)Step 2, Epot=-3.886224e+06, Fnorm=2.500e+01, Fmax=2.186e+02 (atom 7929)Step 3, Epot=-3.886326e+06, Fnorm=1.493e+01, Fmax=1.111e+02 (atom 17398)Step 4, Epot=-3.886397e+06, Fnorm=1.187e+01, Fmax=1.849e+02 (atom 21160)Step 5, Epot=-3.886454e+06, Fnorm=1.538e+01, Fmax=6.140e+02 (atom 21160)Step 6, Epot=-3.886506e+06, Fnorm=3.528e+01, Fmax=2.158e+03 (atom 21160)Step 7, Epot=-3.886618e+06, Fnorm=8.666e+00, Fmax=3.336e+02 (atom 21160)Step 8, Epot=-3.886640e+06, Fnorm=7.111e+00, Fmax=2.319e+02 (atom 21160)Step 9, Epot=-3.886656e+06, Fnorm=1.163e+01, Fmax=6.900e+02 (atom 21160)Step 10, Epot=-3.886692e+06, Fnorm=1.230e+01, Fmax=4.283e+02 (atom 21160)Step 11, Epot=-3.886717e+06, Fnorm=7.702e+00, Fmax=3.192e+02 (atom 21160)Step 12, Epot=-3.886743e+06, Fnorm=6.654e+00, Fmax=9.789e+01 (atom 14750)
writing lowest energy coordinates.
Polak-Ribiere Conjugate Gradients converged to Fmax < 100 in 12 stepsPotential Energy  = -3.8867428e+06Maximum force     =  9.7894951e+01 on atom 14750Norm of force     =  6.6541429e+00
Simulation ended prematurely, no performance report will be written.
So I think this is a problem of emtol in my system, shouldn't be very low, right? 
Kind Regards,Ahmed Mashaly