[gmx-users] overcome the integration problem (installing CVS)

[Michael Shirts]

Hi, all-

> I would like to add that for a constant temperature ensemble the effect
> of this fix is neglegible.
> It is only useful for constant NVE simulations.
>
> Berk.

I think this might need some qualifications. If the constant
temperature dynamics are based inherently on NVE dynamics, they must
be affected similarly.  For example, for Nose-Hoover thermostats,
there exists a conserved quantity.  I would expect that this fix would
also improve that conserved quantity, wouldn't it?

One feature that would be useful is that for any algorithm that has a
"conserved energy-like" quantity, there should be an entry in the .edr
file with that quantity.  It would be a useful check for algorithmic
consistency.  I don't think this is yet in GROMACS, though I could be
wrong -- I haven't examined the CVS that recently.

Best,
Michael Shirts
Research Fellow
Department of Chemistry
Columbia University