Moving integer division to floating-point is trivial