12.1.4.3 Rule of Float Underflow and Overflow

An error of type floating-point-overflow or floating-point-underflow should be signaled if a floating-point computation causes exponent overflow or underflow, respectively.