[gmx-users] Re: System size and minimum image convention

[Dr. Vitaly Chaban]

> Hello everyone,
>
> Does anyone know if the minimum image convention has something to do with
> the box size effect (independence of simulation results from system size) ?
> i.e. When the box size is larger than 2*r_c, does this ensure that the
> results are independent of size pf the system (number of atoms in box)?
>
> Please guide me if I am confusing these two issues with each other,


They have nothing to do with one another.

When the box size is larger than 2*r_c, this ensures nothing, except
this is a [mathematically] correct setup.

You can unveil the box size effect only by running a series of
simulations with different system sizes.

In practice, if your system consists of more than 10,000 interaction
sites and is isotropic, you can forget about box size problem(s). Of
course, it also much depends on the property you are calculating.


Dr. Vitaly Chaban