real(kind=dp) :: fine_gmax

The fine grid will be chosen to be of the minimum size such that all g-vectors with $\vert\bf g\vert \leq {\tt fine\_gmax}$ are included in the fine grid.

If not specified, fine_gmax will be set to -1 ${\rm a_0}^{-1}$, resulting in the fine and standard grids being identical. In this case, the correct value of fine_gmax will be set by the subroutine basis_initialise.