[gmx-users] Re: density problems when switching from md to sd integrator

[Aymeric Naômé]

I have changed tau_t=0.2 for 1.0 but the density is not restored. I have tried with the parameters from Villa and Mark (J Comput Chem, 2002, 23, 548)
, tau_t=0.2 and tau_p=2.0 and still get the same density. The self-diffusion coefficient is indeed half of that with md with sd and tau_t=0.2 (0.71 vs. 1.48 10-5 cm2/s)
and about the same with tau_t=1.0 (1.57 10-5 cm2/s). Water is less affected (1009 g/L with md and 1007 g/L with sd and tau_t=0.2). What is the reason why the density is lowered?

Here is the .mdp file (for the simulation with md, tau_t=0.1):

title		    =  cyclohexane
cpp                 =  /lib/cpp -traditional 
define              =  -DFLEXIBLE
constraints         =  none
integrator          =  sd
dt                  =  0.001	; fs !
nsteps              =  2000000	; total 2000 ps.
nstcomm             =  1
nstxout             =  1000
nstvout             =  1000
nstfout             =  0
nstlog              =  500
nstenergy           =  500
nstlist             =  10 
ns_type             =  grid
rlist               =  1.0
coulombtype		=  PME
rcoulomb            =  1.0
rvdw                =  1.4
fourierspacing		=  0.12
fourier_nx		=  0
fourier_ny		=  0
fourier_nz		=  0
pme_order		=  6
ewald_rtol		=  1e-5
optimize_fft		=  yes
energygrps		=  system
; Berendsen temperature coupling is on in three groups
Tcoupl              =  berendsen
tau_t               = 0.2    
tc_grps		    = System 
ref_t               = 298    
; Pressure coupling is on
Pcoupl              =  berendsen
pcoupltype          =  isotropic
tau_p               =  0.5
compressibility     =  11.2e-5
ref_p               =  1.0
; Generate velocites is on at 300 K.
gen_vel             =  no
gen_temp            =  298.0
gen_seed            =  173529
; Energy group exclusion
energygrp_excl      =  
freezegrps          =  
freezedim           =  
; Non-equilibrium Thermodynamics
acc_grps	    =  
accelarate	    =  
; center of mass
comm_mode           = 
comm_grps           =

> Dear Aymeric:
>
> 1. Can we please see the entire .mdp files for both simulations? I  
> suggest that you use a tau_t=1.0 (0.2 is probably over-damped).
>
> 2. Although any value of tau_t should still produce the correct  
> equilibria, your diffusion rates and your overall sampling may be  
> slower with sd tau_t=0.2 than they are with md.
>
> 3. Can you reproduce this effect with a box of water?
>
>> -- original message --
>>
>> I'm simulating a box of cyclohexane (GROMOS 53a6, 512 molecules, 298K,
>> compressibility=11.2 10-5 bar-,1 tau_p=0.5 ps, no constraints on bonds).
>> After equilibration and NPT simulation, the system reaches the proper
>> density reported for the model (791g/L). When I switch to a stochastic
>> integrator (with inverse friction coeff. tau_t=0.2 ps) the density
>> rapidly falls to about 765g/L, the potential energy increases, while
>> temperature and pressure fluctuate much more around their specified
>> values. I intend to use this box of cyclohexane for solvation free
>> energy calculations. Can anyone explain what is going on?
>> Thanks in advance
>>
>> Aymeric

-------------- next part --------------
A non-text attachment was scrubbed...
Name: anaome.vcf
Type: text/x-vcard
Size: 318 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100201/eb2e2a33/attachment.vcf>