[gmx-users] Re: Re: Energy conservation of rigid water models

[David van der Spoel]

Janne Hirvi wrote:
>>Have you compiled gromacs in single precision?
>>I think double precision compilation is critical
>>for energy conservation in NVE.
> 
> 
> Yes, I have used only single precision compilation because it has been enough
> accurate until now (NPT/NVT). Actually I just want to know if employed time
> step of 2fs is applicable and so I think that I should use same compilation
> also for energy conservation simulation in NVE ensemble. Or may I prove
> suitability of this long time step with double precision compilation even I
> have done other simulations with single precision compilation?
> 

It is important to realize that you need this kind of precision only in 
special cases, for instance when you want to derive properties from the 
fluctuations of T, P, E.

The algorithms in GROMACS are not perfect, that is, there are better 
algroithms for integrators etc. We are working on these things.

However, what is implemented, i.e. leap-frog, when used carefully, does 
conserve energy. Things that spoil energy conservation are:

- T coupling
- P coupling
- Cut-offs (both LJ and Coulomb)
- Constraints (but not SETTLE as used for water)
- Long time steps (in relation to the fastest vibrations)
- Too long intervals between neighborsearching
- To some extent, single precision (due to the fact that you sometimes 
add small forces to large one, and by this lose precision).


-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++