floor() returns the largest floating-point number of integral
    value less than or equal to x.

                           Entry-Point Names

    Generic                               Compaq
    Function   Data Type   OpenVMS        Tru64 UNIX
    Name       Required    Alpha          Alpha

    floor      S_FLOAT     math$floor_s   floorf
               T_FLOAT     math$floor_t   floor
               X_FLOAT     math$floor_x   floorl
               F_FLOAT     math$floor_f
               G_FLOAT     math$floor_g