[gmx-users] how to keep the simulation box during simulation

[Itamar Kass]

  Hi all, I am trying to simulate a lipid bilayer (at plain XY, such that it is normal to Z) with water at the two sides.  When I do PR dynamics, using:

title               =  DMPC bi-layer solvated in water and Na ions.
cpp                 =  /usr/bin/cpp

integrator          =  md
dt                  =  0.002
nsteps              =  50000

; Constrain control
define               = -DPOSRES
constraints          = all-bonds

; Output frequency for coords (x), velocities (v) and forces (f)
nstxout             =  500
nstvout             =  500
nstfout             =  500

; nblist update frequency
nstlist             =  10 

; ns algorithm 
ns_type             =  grid
rlist               =  1.3
coulombtype	    =  PME
rcoulomb            =  1.3
rvdw                =  1.3
fourierspacing	    =  0.10
fourier_nx	    =  0
fourier_ny	    =  0
fourier_nz	    =  0
pme_order	    =  6
ewald_rtol	    =  1e-5
optimize_fft        =  yes

; Berendsen temperature coupling is on in three groups
Tcoupl              =  berendsen
tau_t               =  0.1      0.1  
tc-grps		    =  Solvent  DMPC
ref_t               =  310      310 

; Pressure coupling is on
Pcoupl              =  berendsen
pcoupltype          =  semiisotropic
tau_p               =  2 2
compressibility     =  4.5e-5 0
ref_p               =  1.0 1.0

; Generate velocites is on at 310 K.
gen_vel             =  yes
gen_temp            =  310

; Electric field - used
; 100mV/3nm*10E-3v/mV = 0.033
E_z = 1  0.033  0.0

; Energy monitoring
energygrps          =  SOL DMPC

; Center of mass control
nstcomm             =  1
comm-grps           =  Solvent  DMPC

I got from the mdrun a box with different size then the one entered.  I know I used pressure coupling, which might change the box, but the bilayer is ruined. I saw in articles (e.g. Biophys J. 2002 Jun;82(6):2934-42.) that it is possible to allow changes only at the z axis, how this is done?  Does working this way, will keep the bilayer and it's properties physically?

  And a second question if I may, when running gromacs, is it better to give to any CPU more then one job.  For instance, if I have two CPU is it more efficient to cut the job into four parallel processes?

     Thanks you all, Itamar.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20031216/c0797c1b/attachment.html>