[gmx-developers] Re: Welcome to the "gmx-developers" mailing list

JMandumpal jesbman at rediffmail.com
Mon Aug 11 05:58:08 CEST 2008

  Dear GROMACS developers,

                               In order to measure the glass transition temperature of water and aqueous solutions with different cooling rates,   Stanley et al have reported a modified NVT simulation - Physical Review Letters 93, 2004 047801-1.

His method is as follows:

The simulation is run in the NVT ensemble. The temperature during the simulation, ref_T, is controlled by using a Berendsen thermostat . 

In a standard NVT simulation, To is constant. In their cooling/heating simulations,  Berendsen thermostat is used but, at everytime step, they also update the bath temperature ref_T such that 

ref_T= ref_T - deltaT, with deltaT= cooling/heating rate *  timestep.

This is done at every time step before rescaling the velocities.

In my case, I want to employ a cooling rate of -3*10^10 K/fs and time step is 1 fs.

that is delta T is -0.00003K/fs

and ref_T has to be updated at everytime step using the equation,

ref_T= ref_T - 0.00003 -----------> eq.1

How can I do this in GROMACS 3.3.3?

I mean which file should I modify before compiling again;

I checked md.c, do_gt.c and coupling.c  in 3.3.3/src/kernel but got confused as to find the line where I could impliment eq.1.

I really hope that any of you can help me in this regard.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20080811/1696c2aa/attachment.html>

More information about the gromacs.org_gmx-developers mailing list