[gmx-users] User-defined potentials and meaning of table extension

[Mark Abraham]

With the group scheme, rlist constructs a neighbour list that is updated
every nstlist steps. The relative size of rlist and (say) rcoulomb
determines what that means, e.g. for having a buffer. If there is no
potential modifier, then rcoulomb is not used in the kernel, so particles
can diffuse past rcoulomb over the lifetime of the list. Then, a table
extension might be wanted. But I think the need for it is anachronistic
(from when there were no modifiers).

Buffering, strict cutoff, and modifier are all automatic in the verlet
scheme. AFAIK table-extension is then pointless.

Your water experiment should work with zero table-extension and nstlist=1.
Please let us know how you go!

Mark
On Mar 21, 2014 10:39 AM, "Davide Mercadante" <dmer018 at aucklanduni.ac.nz>
wrote:

> Hello GMX users,
>
> I am testing the use of User-defined on a box of water and I came across
> the issue of table-extension option.
>
> I don't understand the use of it. If I specify a cutoff of 1 nm and then
> give a table extension of 0 nm then the cutoff will be 1 nm otherwise it
> will be cutoff+table-extension value. This is pretty clear. However, it is
> not clear why a table extension is needed.
>
> As a test, I have run this on only two water molecules placed at a distance
> within the cutoff coupled to table-extension = 0nm and to table-extension =
> 1 nm.
> In both cases, the values listed in the table seem to be considered up to
> the cut-off and the table extension seem to be ignored?
> A puzzling thing that happened is that then, for a box of water (216
> molecules), using table extension = 0 nm (cutoff 0.5 nm) I get an unstable
> and crashing simulation, while if I use table extension = 1 nm (cutoff 0.5
> nm) the system runs smoothly.
>
> Can you please help me to understand this?
>
> Any comments will be very much appreciated.
>
> Thank you in advance for your time.
>
> Davide
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>