[gmx-users] energy drift in nve with conservative parameters

[Erik Lindahl]

On Tuesday, Mar 4, 2003, at 11:50 US/Pacific, Claudio J. Margulis wrote:

> I am sure I am not using constraints.
> I attached, the mdp file in the first email I sent a little while ago.
>
> I am not quite convinced that single precision is the problem. Any 
> insight on more stringent pme parameters?
>

Ah, sorry, I missed that file and the FLEX_SPC option.

You could try to decrease ewald_rtol to 1e-7 or so - this will get the 
direct-space potential closer to zero at the cutoff, but you might have 
to decrease the fourierspacing too then.

Cheers,

Erik