[gmx-users] Re: g_energy average --- bug?

[Berk Hess]

>From: "pim lists" <pimlists at googlemail.com>
>Reply-To: pimlists at gmail.com,Discussion list for GROMACS users 
><gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] Re: g_energy average --- bug?
>Date: Thu, 2 Mar 2006 13:59:48 +0100
>
>From: pim lists <pimlists at googlemail.com>
>Date: Mar 2, 2006 12:17 PM
>Subject: Re: g_energy average --- bug?
>To: gmx-users at gromacs.org
>
>Date: Thu, 2 Mar 2006 10:49:38 +0100
>
> > From: "Fernando Mattio" < mattiofer at gmail.com>
> > Subject: [gmx-users] g_energy average --- bug?
> >
>
>    However, the average that I got from g_energy was 264,694 K. I really 
>can
>
> > not understand why this difference, it should be some value between 261 
>K
> > and 263,5K. Is it a bug from g_energy? Should I trust in this average 
>even
> >
> > knowing that the temperature accordind to the temperature file didn´t
> > reached this value?? I wait for kind informations.
>
>
>As far as I know g_energy averages and rmsd values that are given in stdout
>are always a bit off. You should only use those for rough information. It
>gets even worse when you have separate ener.edr files that you look at in a
>row.
>
>You still can get correct data though: to get those, perform g_analyze over
>the resulting energy.xvg file. In any case g_analyze will give you more
>options to analyze the data ( e.g. test for error estimates etc). The

Wrong!

This issue has come up many times on the mailing list.

Data for exact averages using every MD time step are stored in
the energy file. The average given by g_energy is therefore
always more accurate than that obtained by averaging the values
in the output xvg file (unless you used nstenergy=1).

Berk.