[gmx-users] How to add another electrostatic summation method in Gromacs

[Mark Abraham]

Shuangxing Dai wrote:
> Hi, all,
>     I was wondering how to add another electrostatic summation 
> method--Wolf summation into Gromacs. Wolf summation is a treatment for 
> summation of Coulombic energy. Since the formula of total potential 
> energy for each ion is known, how to add this into Gromacs? According to 
> Chapter 7, I should give the f(x) and f'(x). The formula of f and f' 
> depends on the neighbours and varies from ion to ion, and I can not give 
> f and f' without calculation.

If you have a functional form that depends only on pairwise distance and 
two atom types, then you can use the GROMACS table lookup machinery.

> Since Wolf method uses the same error 
> function as Ewald summation, it might be easier and possible to 
> implement Wolf into Gromacs. If so, which files need to be modified?
> Hope I put my question clearly. Thank you in advance for suggestions and 
> help.

Evaluating erfc is not the core issue - but GROMACS does the evaluation 
of terms for the direct-space Ewald contribution via table-lookup. The 
big deal is whether you need a custom non-bonded kernel, and then 
whether you can make it run fast. You might need to post a link to some 
details of this summation method to get useful help.

Mark