[gmx-users] heat exchanges

[chris.neale at utoronto.ca]

> The integrations errors in the electrostatics have much more effect
> on the water than on the protein, because the water has higher charges
> and is far more mobile. No thermostat can correct for these errors,
> unless you use multiple groups.

I realize that it is not exactly a "thermostat", but the SD integrator  
is expected to correct for these errors using only one temperature  
coupling group is it not?