Optimisation

Clearly a better solution to the problem of the projector-overlaps in the Ion module would be to distribute the $\beta$-projectors as well as the $\gamma$-projectors. However the $\beta$-projectors are used extensively throughout the Ion module, including computing overlaps between wavefunctions and the projectors which can be done without communication in the current code. Distributing the $\beta$-projectors would thus require extensive modifications and an extra communication overhead, whereas the changes described above were localised to the problematic subroutine.