[gmx-users] Question about Berendsen thermostat and Nose-Hoover temp coupling

[David van der Spoel]

David Osguthorpe wrote:
> On Tue, Jul 22, 2008 at 11:16:21PM -0400, chris.neale at utoronto.ca wrote:
>> Thanks David for sharing your knowledge, especially the note that for  
>> further information one can refer to the literature that was written  
>> around the time that these thermostats were released. I have a  
>> question though:
> 
> I was just trying to correct the impression from the posts that the issue with
> Berendsen is some urban legend - it is not and there is extensive
> literature on it - even if its not googleable
> 
>>> by the way it is also true that if you use a thermostat or barostat  
>>> then although long time averages are equivalent to averages in the  
>>> NVT or NPT ensemble strictly the dynamics is no longer valid as
>>> a Newtonian trajectory so you should not derive dynamic properties  
>> >from such trajectories.
>>
>> Have you seen any information to suggest that this is actually a  
>> non-trivial concern? That is, given static point charges, an empirical  
>> LJ force, short cutoffs, etc., do you believe that the application of  
>> nose-hoover, berendsen, or even the arbitrary velocity rescaling  
>> significantly degrades the quality of the obtained dynamics?
>>
> 
> I think the response by Michael says it as well as I could - this was a
> point made by some of these papers - so just because the averages are
> valid does not mean the dynamics is valid - you are right in that it
> may not be something that has an observable effect with current simulations
> but it is some form of artifact that may bias the simulations consistently
> whereas eg. with force fields you can have lots of cancellation of errors
> 
Cancellation of errors is not an excuse for poor force fields. In 
addition, built-in cut-offs for some force fields also give systematic 
bias, but this is beside the point.

Although it is indeed well-known that Berendsen thermostat (and 
Barostat!) create significant artifacts, these are not as bad as they 
used to be, in particular due to the use of PME. Temperature coupling 
hardly does anything at all when using PME, which you can verify by 
plotting the temperature scaling factor as a function of time.

So we are all waiting for a healthy volunteer to implement the latest 
and greatest algorithms in the latest and greatest gromacs development code.

Cheers,
-- 
David van der Spoel, Ph.D., Professor of Biology
Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala University.

Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se