[gmx-users] Different Coulomb energy between gromacs and my calculation

[Tanping Li]

Dear Gromacs users,

When I choose cutoff in Coulomb term, in my
understanding, if one pair is within cutoff, I can get
the exactly same result as Gromacs. I use formular:

e(i,j)=1389*q(i)*q(j)/r.


But my test result from one charge group and one Na is
6.250759; in gromacs, I set the step as 0, the result
is -6.252050.

It does not look like rundoff. This difference can be
enlarged in my real calculation between a certain
residue and waters.

How does gromacs calculate the coulomb interaction in
cutoff? Does it calculate between individual pairs and
fianlly get the sum?

Thank you very much. I tried all ways. It is really
hard to find the solution.


Tanping