Re: double-wide sampling with soft-core

From: Jerome Henin (
Date: Wed Jan 21 2009 - 16:02:24 CST

Hi Sébastien,

On Tue, Jan 20, 2009 at 3:27 PM, Sébastien Légaré
<> wrote:
> Dear NAMD users and developpers,
> I have a theoretical question about the use of separation-shifted vdw in FEP,
> now available in the cvs code. As stated in the FEP tutorial, linear scaling
> of the potential allows double-wide calculation with a single transformation.
> The only way I found to do this precisely is to post-process with all computed
> dE, so I need to set fepOutFreq to 1. Is there an other way?

No, formally this is correct. Numerically though, samples of dE are
correlated, so you get pretty much the same statistics by storing it
every 10 timesteps or so. You could try it - downsampling dE data from
a calculation, and seeing how far you can go without changing the

> My main question is: Since it is not a linear scaling, shouldn't it be
> necessary to provide a second comparison lambda in order to do double-wide
> calculations with separation-shifted vdw?

Again, you are right: actually this is now implemented. The problem
is, I think, that the user's guide in CVS has not yet been updated to
document this. There is some work being done on that side...

> Thanks for this implementation, it will be very useful.

I am thankful to the people who did it, too :-)


This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:52:16 CST