[gmx-users] "ideal" mdp file for NVE simulation as textbook example

[Marc Baaden]

Hi,

I am currently preparing/improving a tutorial on MD which is based
on Gromacs. One of the things would be to compare a "normal" run
to a run where all is conserved as good as possible in an NVE
ensemble, that is the energy should be as constant as possible.
For this I have tried the attached mdp file, with infinite cutoff,
no constraints and no P-/T-coupling. I run it in double precision.

Still the energy goes down (ca 4 kJ per 10 ps) with increasing
noise. (See http://www.shaman.ibpc.fr/nvetest.png at the bottom)

Now I wonder whether this is as good as it gets, or whether I
overlooked some other option to avoid numerical and algorithmic
inaccuracies ?
(I can't run PME, I am in vacuo and charged; the system is bpti)

Thanks for any suggestions,
  Marc Baaden


-------------- next part --------------
;
;	NVE test run
;
title               =  Dynamique de BPTI
cpp                 =  /lib/cpp
constraints         =  none
integrator          =  md
dt                  =  0.001	; ps !
nsteps              =  10000	; total 10 ps.
nstcomm             =  -1
comm-mode           =  ANGULAR
nstxout             =  50
nstvout             =  1000
nstfout             =  0
nstlog              =  100
nstenergy           =  100
nstlist             =  1
ns_type             =  simple
rlist               =  0.0
rcoulomb            =  0.0
rvdw                =  0.0
; Berendsen temperature coupling is on in two groups
Tcoupl              =  no
;Tcoupl              =  berendsen
tc-grps		    =  
tau_t               =
ref_t               =
; Energy monitoring
energygrps          =  Protein
; Isotropic pressure coupling is now on
Pcoupl              =  no
; Generate velocites is off at 300 K.
gen_vel             =  no
gen_temp            =  0.0
gen_seed            =  173529

pbc = no
-------------- next part --------------
 BioMolSim meeting 2&3 Sep 2005: http://www.iecb.u-bordeaux.fr/satellite2005/

 Dr. Marc Baaden  - Institut de Biologie Physico-Chimique, Paris
 mailto:baaden at smplinux.de      -      http://www.baaden.ibpc.fr
 FAX: +33 15841 5026  -  Tel: +33 15841 5176  ou  +33 609 843217