Returns a
vector containing the smallest representable floating-point integral
values greater than or equal to the values of the corresponding elements
of the given vector.
Note: vec_ceil provides
the same functionality as vec_roundp, except that vec_ceil could
trigger the IEEE-inexact exception.