[gmx-users] Steepest Descents converged to machine precision in 7559 steps, but did not reach the requested Fmax < 100.

[Justin Lemkul]

On 1/27/13 11:53 PM, Ewaru wrote:
> Dear Gromacs experts,
>
> I am running gromacs for the first time, and this is my em.mdp file.
>
> define = -DFLEXIBLE
> constraints = none
> integrator = steep
> dt = 0.01 ; ps !
> nstlist = 10
> ns_type = grid
> rlist = 1.0
> coulombtype = PME
> rcoulomb = 1.0
> vdwtype = cut-off
> rvdw = 1.4
> fourierspacing = 0.12
> fourier_nx = 0
> fourier_ny = 0
> fourier_nz = 0
> pme_order = 4
> ewald_rtol = 1e-5
> optimize_fft = yes
> DispCorr = EnerPres
> nsteps                   = 10000
> nstenergy       = 1000
> nstxtcout       = 1000
> ;
> ; Energy minimizing stuff
> ;
> emtol = 100.0
> emstep = 0.01
>
>
> Unfortunately, when running mdrun, I obtained this "Steepest Descents
> converged to machine precision in 7559 steps, but did not reach the
> requested Fmax < 100."
>
> Please, can somebody tell me what do I need to change in order for it to run
> until the requested steps? Also, the RMSD graph obtained does not begins
> from zero (how can I make it to?) and it does not converge.
>

http://www.gromacs.org/Documentation/Errors#Stepsize_too_small.2c_or_no_change_in_energy._Converged_to_machine_precision.2c_but_not_to_the_requested_precision

Regarding the RMSD, it probably won't converge during EM since the structure 
changes at every step.  Without seeing your actual command and the snippet of 
the .xvg file showing the starting value, it's impossible to say why it doesn't 
start at zero.  If it is a very small value (i.e. 1e-4 or less) and not exactly 
zero, you can consider that to be effectively zero.  Just one of the quirks of 
floating-point math.

-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

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