logb() returns a signed integer converted to double-precision
    floating-point and so chosen that 1 <= |x|/2**n < 2 unless x = 0
    or |x| = infinity.

    IEEE Std 754 defines logb(+infinity) = +infinity and logb(0) =
    -infinity. The latter is required to signal division by zero.

                           Entry-Point Names

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

    logb       S_FLOAT     math$logb_s    logbf
               T_FLOAT     math$logb_t    logb
               X_FLOAT     math$logb_x    logbl
               F_FLOAT     math$logb_f
               G_FLOAT     math$logb_g