Vectorization of Cosmological Gadget-2 interaction functions

With the introduction of the cosmological time-integration, a few additional terms have to be computed in the force interaction routine for SPH. They have been added to the scalar version but not yet to the vectorized version.

The difference is that fac_mu is no longer a constant and that vdotr receives an additional term. Both these things can be computed from the value of a and H extracted from the cosmology structure.

It'd be nice to add these terms as well to the vectorized routines.