[gmx-users] Berendsen Barostat for production runs

[Daniel Bauer]

Hi,

I know that the usage of Berendsen baraostat for production runs is
generally discouraged because it does not give a correct ensemble
average. Nevertheless, I found some recent publications where this is
done (presumably for performance reasons). Example: 10.1126/science.1254840

What would be a justification that allows to do this? Under what
conditions is it "ok" and when is it not? An explanation attempt by me
is that Berendsen is fine if we do not try to investigate properties of
the system that depend on the ensemble average (i.e ion conduction events).

Best wishes,

Daniel

-- 
Daniel Bauer, M.Sc.

TU Darmstadt
Computational Biology & Simulation
Schnittspahnstr. 2
64287 Darmstadt
bauer at bio.tu-darmstadt.de

Don't trust atoms, they make up everything.