[gmx-developers] [gmx-users] Wrong Nose-Hoover integrator

[David van der Spoel]

Michel Cuendet wrote:
> 
> Hi,
> 
> To maybe conclude this old thread about calculating the temperature
> within a leap-frog scheme, a little paper just came out :-)
> 
> http://link.aip.org/link/?JCPSA6/127/184102/1
> 
Congratulations.

Berk implemented this in GROMACS a few years ago.

	/* Calculate the full step Ekin as the average of the half steps */
	for(j=0; (j<DIM); j++)
	  for(m=0; (m<DIM); m++)
	    tcstat->ekin[j][m] =
	      0.5*(tcstat->ekinh[j][m] + tcstat->ekinh_old[j][m]);
       }

> Cheers,
> Michel
> 
> 
> Berk Hess wrote:
>> If the (say correct) half step temperature of 300 K gives a full  step 
>> temperature
>> of 297.8 K, this means that the actual temperature of all water  
>> simulations
>> done with Gromacs is 2.2 K too high as we couple the full step  
>> temperature
>> to 300 K. It is even worse for my 4 fs water runs, which are 9 K  too 
>> high.
> 


-- 
David van der Spoel, Ph.D.
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