mlund / coulombgalore Goto Github PK

If I get it correctly, with the current definition of r, the function returns the force acting on charge "b". Perhaps we could change to @brief force on charge a due to charge b? Due to a lucky cancellation of signs, we got the right force in Faunus, but it's quite easy to make mistakes. I renamed distance vectors to a_towards_b to make it very explicit what is expected.

Add unittests for self-energies

Add neutralising background for charged systems

We should define a policy for the units for function arguments and return values. Currently there are suggestions for Hartree units in the comments(?)

ping @bjornstenqvist

Split reciprocal ewald part into a EwaldData object as in Faunus. The functions for calculating the surface and reciprocal energy should not be part of the base class (as they are specific to Ewald). Functions for updating when one or many particles have moved should be added.

ping @bjornstenqvist