Implement tail corrections to the energy and virial

[Luthaf]

The core logic is implemented here for energy and here for the virial, using the PairInteraction::tail_energy and PairInteraction::tail_virial functions.

This need a bit more tests, @g-bauer what was your strategy for checking against NIST data?

Fix #28

[g-bauer]

This need a bit more tests, @g-bauer what was your strategy for checking against NIST data?

I use the NIST reference calculations. I build an example, where the configurations offered by NIST are read and the system energy is calculated and checked vs. the values from NIST. I used the examples to test the energies without tail corrections. I'll send you the examples.

Regarding the PR, I think the name tail_virial is a bit misleading since you only use it to compute a correction to the pressure.

[Luthaf]

Regarding the PR, I think the name tail_virial is a bit misleading since you only use it to compute a correction to the pressure.

It is used in the computation of the Virial tensor, which in turn is used for the pressure.

I agree that the name Virial for this tensor is bad. It is more something like the virial contribution to the stress tensor. I'd like to improve the name, but I have no idea of a better one for the moment.

[Luthaf]

I added the test comparing against NIST, but a kind of potential is missing to compare with the LFS column.

Linear Force-shift potential can be added later (and are actually an easy thing to do), see issue #55 for this.