float-digits
(float-digits float)
Returns the number of radix-2 digits the representation of float uses, including the implicit leading bit: 53 for every float, since every float is a double. The digits a value actually holds -- fewer for a subnormal, none for a zero -- are float-precision's answer.