[gmx-developers] Position-dependent tabulated nonbonded potential tables for a single energy group pair

[Berk Hess]

Using two different potentials will cause all kinds of trouble, as you 
will have jumps
in the potential when a particle moves from one region to the other.
If you want to do something like this you have to come up with a potential
which depends continuously on the distance and z-coordinates of the two 
particles.
You could implement this in the generic non-bonded kernel.

Cheers,

Berk

On 03/28/2012 01:33 AM, Beste Bayramoglu wrote:
> Thanks for your response Anton!
>
> The reason that I want to use different potentials for the same pairs 
> in different regions is that because they are in different 
> environments when they are in different regions. I cannot use the same 
> potentials for all the environments they are in, so that's why I 
> should modify the code such that it assigns a specific potential table 
> for, let's say 0<box_z<2nm, and then another one for the rest of the 
> simulation box_z (for the same energy group pair, of course).
>
> Any comment from anyone is appreciated.
>
> Thanks,
>
> Beste
>
> On Tue, Mar 27, 2012 at 5:05 AM, Anton Feenstra <k.a.feenstra at vu.nl 
> <mailto:k.a.feenstra at vu.nl>> wrote:
>
>     On 27/03/12 10:30, Beste Bayramoglu wrote:
>
>         Hi,
>
>         I've been performing molecular dynamics simulations using
>         tabulated
>         potentials with Gromacs version 4.0.7. Everything works fine
>         with one
>         tabulated potential table for each type of energy group pair.
>         However,
>         what I need to do now is that I need to use multiple
>         position-dependent
>         nonbonded potentials (in tabulated form) for each of my energy
>         group
>         pairs.That's to say, for a certain energy group pair, I need
>         to assign
>
>         multiple tables and the program should decide on which one to
>         use in the
>         calculations depending on the positions of the atoms along
>         z-direction.I
>         was wondering if any unofficial attempts to modify the source
>         code have
>         been made so far for a similar purpose. And which part of the code
>         should be changed? I checked the mail list archives, but could
>         not find
>         such an entry. I know that this is going to be a very
>         difficult task for
>         a newbie like me and probably would take quite some time, so
>         I'd like to
>         know beforehand if what I want to do is reasonable.
>         Unfortunately, I
>         don't have much time until I graduate, so if it's not
>         reasonable I'd
>         have to think about other alternatives. I would really
>         appreciate if
>         anyone could help me with this.
>
>
>     This sounds very complicated.
>
>     It would be good if you could explain what it is that you want to
>     achieve with these interactions?
>     That way you may benefit from the combined expertise of the list
>     members in finding a workable solution for your problem.
>
>
>     -- 
>     Groetjes,
>
>     Anton
>      _____________ _______________________________________________________
>     |             |                                                  
>         |
>     |  _   _  ___,| K. Anton Feenstra                                
>         |
>     | / \ / \'| | | IBIVU/Bioinformatics - Free University  Amsterdam
>         |
>     |(   |   )| | | De Boelelaan 1083A - 1081 HV Amsterdam -
>     Netherlands  |
>     | \_/ \_/ | | | Tel +31 20 59 87783 <tel:%2B31%2020%2059%2087783>
>     - Fax +31 20 59 87653 <tel:%2B31%2020%2059%2087653> - Room P136 |
>     |             | Feenstra at few.vu.nl <mailto:Feenstra at few.vu.nl> -
>     www.few.vu.nl/~feenstra/ <http://www.few.vu.nl/%7Efeenstra/>         |
>     |             | "Carbohydrates is all they groove" (Frank Zappa)  
>        |
>     |_____________|_______________________________________________________|
>
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20120328/41dee0e2/attachment.html>