[gmx-users] heat exchanges

[Berk Hess]

Hi,

With SD there are no temperature differences to be expected when the total
temperature is correct.
It could only go wrong in the very rare case that one group heats due to cut-off
effects and another one cools due to inaccurate constraints.
But because you get the same temperature at both tau_t's you can be certain
there are no problems.

The energy fluctuation issue is strange, I thought this was all fixed by now.
But maybe you use a strange continuation procedure.

Berk

> Date: Thu, 5 Feb 2009 15:08:26 -0500
> From: chris.neale at utoronto.ca
> To: gmx-users at gromacs.org
> Subject: [gmx-users] heat exchanges
> 
> Thank you Berk,
> 
> I have checked my temperature and find that there is no difference while 
> using tau_t=0.1 or 1.0 (~2 ns):
> tau_t=0.1: temp=303.36 +/- 1.90
> tau_t=1.0: temp=303.31 +/- 1.98
> 
> While using only one temperature coupling group, how would I ensure that 
> the temperature of water is
> the same as that of DPC?
> 
> I do see the large depression of diffusion that is induced by tau_t=0.1 
> and is pointed out in your thesis:
> 
> Diffusion of tip4p:
> tau_t=0.1: 1.43 +/- 0.003
> tau_t=1.0: 2.81 +/- 0.136
> md with berendsen = 3.115 +/- 0.056
> 
> Diffusion of DPC detergent:
> tau_t=0.1: 0.0323 +/- 0.003
> tau_t=1.0: 0.137 +/- 0.02
> md with berendsen = 0.0928 +/- 0.045
> 
> PS: Note that the RMSD values for the temperature  above are from my own 
> calculation as those from g_energy
> in gmx 3.3.1 and 4.0.3 are nonsensical to me :
> 
> Energy                      Average       RMSD     Fluct.      Drift  
> Tot-Drift
> -------------------------------------------------------------------------------
> Temperature                   303.3    474.854    474.854 -3.75465e-05 
> -0.0976211
> 
> probably because thie .edr was generated by eneconv on all my 200 ps 
> segments.
> Indeed, g_energy on a single segment gives me a sensible RMSD value.
> 
> Thanks again,
> Chris.
> 
> #####################
> -- original message --
> 
> Yes, the SD "thermostat" has the huge advantage that it thermalizes each
> degree of freedom separately.Therefore you never have incorrect temperature
> distributions in your system, unless you have very serious integration
> artifacts. SD also avoid any ergodicity problems.
> 
> BTW You should check your temperature for your SD runs with tau_t=1.
> I often use 1, but that is with a shifted LJ potential and not with a 
> twin-range
> setup that introduces some errors. You might need to change to tau_t=0.5.
> tau_t=1 has nearly no effect on the dynamics of water, whereas tau_t=0.5
> slows it down somewhat. See chapter 2 of my thesis for numbers for SPC
> (which diffuses twice as fast as real water):
> http://dissertations.ub.rug.nl/faculties/science/2002/b.hess/
> 
> Berk
> 
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090205/1a35709c/attachment.html>