Double precision floating point value which is nearest to this number. Uses the same semantics as conversion from ulong to double.
See Implementation