[gmx-users] Re: the Pressure control in GROMACS 4.0.7

[David van der Spoel]

On 2010-07-10 09.47, Vitaly Chaban wrote:
>> Dear users,
>>   I think the pressure control in gromacs 4.0.7 may have some problems. First ,The pressure value>fluctuates widely ,even between -7984.377930 and 5920.053711 bar.Second,the average pressure>sometimes is much larger than what I have set,for example, I have set ref_p  = 1.0 , but the average is>actually over 7 bar. It is because I have not set properly , or something else?
>> Zhongjin He
>
> Hello Zhongjin:
>
> The average pressure in gromacs is really somewhat larger, e.g. if you
> request ref_p=1bar, you always get something close to P=1.3 bar. If
> somebody can comment on this issue, it will be interesting. Anyway, I
> think that the coupling in general is correct, since the simulated
> densities are as they are expected.
>
fluctuations are proportional to 1/sqrt(natoms). The numbers Zhongjin 
mentions are extreme, for a 216 water box *in equilibrium* you would get 
1 +/- 500-600 bar. I suspect your sytem may be out of equilibrium or 
very small. Whether the average is 1 or 5 does not make a difference if 
the fluctuation are two orders of magnitude larger.


> As I suppose, the main reason is that pressure is not a nanoscopic
> property whereas your MD system is nanosized. I would advice you to
> gradually enlarge the system and monitor the pressure. It should give
> you some answers anyway.
>
> The average of  7 bar if one requests 1 bar, is weird. Look at the
> simulation time and the system size firstly.
>
> Best.
>
> --
> Dr. Vitaly Chaban
> Associate Researcher


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se