[gmx-users] rvdw vs cut-off / different types of non bonded interactions

[Mark Abraham]

Claus Valka wrote:
> Hello,
> 

> 3rd question: Which is the difference between the
> S(hort)R(ange) LJ and LJ 1-4? For me the short range
> interactions are the 1-4LJ, yet it seems that this is
> not the case.

GROMACS uses twin-range cutoffs (see manual), so LJ-SR is inside the 
shorter cutoff, and LJ-LR is outside.

> 4th question: I want after 1,45*σ till
> 2,33*σ to use the cubic spline in such a way that
> the LJ potential will go smoothly to zero at
> 2,33σ. If I use the biggest σ I have, then
> in the mdp file rvdw-switch=5.56 and rvdw=8.9. Between
> these two values does the cubic spline work as a
> switch function using vdw-type = Switch ?

Please find a way to describe your problem with plain ASCII text.

> 5th question: Given the above if I use vdw-type
> =Cut-off and rvdw=8.9 then gromacs will use cubic
> spline till the LJ will reach zero for a value greater
> than 8.9, yet to me uknown? If this is not the case
> how LJ goes to zero with this option?

I don't understand what you're asking.

> 6th question: If I want to use a quintic spline
> between 1,45*σ till 2,33*σ(cut-off value)
> for the different combination of atoms(they have
> different ε and σ), then should I use a
> tabulated potential? And if I do that which will be
> the meaning of cutoffs in the mdp file as I would like
> them to be different for each of the different types
> of atoms I have? Can I use different cut-offs for the
> different types of atoms I have?

The cut-offs are a system-wide property.

Mark