# Re: Switching Function

From: Jason O'Young (joyoung_at_uwo.ca)
Date: Thu Sep 20 2007 - 14:41:00 CDT

Thanks Dave!

In that case, what are commonly used switchdist and cutoff parameters?

To my understanding, the GROMACS defaults for the analogous
parameters are 0 and 10 (I am trying to run a GROMACS simulation in
NAMD). As a result, I just chose a small value to replicate it as I
could not use 0. However I am unsure how they handle their switching
function.

Cheers,
Jason

On 20-Sep-07, at 2:39 PM, David Hardy wrote:

> Jason,
>
> The switched potential has the form: U(r) s(r),
>
> where U(r) is your unmodified Lennard-Jones potential
> and s(r) is a smooth, piecewise defined polynomial:
>
> s(r) = 1 for r <= switchdist
> s(r) = 0 for r >= cutoff
> otherwise, s(r) is a decreasing function
>
> So there is no danger of the function being entirely positive,
> although your choice of switchdist will probably give you a much
> shallower well depth than you want.
>
> Regards,
> Dave
>
>
> On Sep 20, 2007, at 1:00 PM, Jason O'Young wrote:
>
>> Hi All,
>>
>> I set my switchdist to 0.01 and my cutoff to 10.
>>
>> I am not too sure how the switching function would look like, but
>> is there a danger of the function being entirely positive?
>>
>> For an example of what I mean, see this picture from the users guide:
>> http://www.ks.uiuc.edu/Research/namd/2.6/ug/img14.png
>>
>> If the switching was to begin at such a low value, would the
>> function just proceed from infinitely positive to 0? Or would the
>> general form of the potential function be "squeezed"?
>>