Function float2half

include/half.hpp:1064–1067

		/// \exception FE_UNDERFLOW on underflows
		/// \exception FE_INEXACT if value had to be rounded
		template<std::float_round_style R,typename T> unsigned int float2half(T value)
		{
			return float2half_impl<R>(value, bool_type<std::numeric_limits<T>::is_iec559&&sizeof(typename bits<T>::type)==sizeof(T)>());
		}

		/// Convert integer to half-precision floating-point.
		/// \tparam R rounding mode to use