An intermediate or final result cannot be represented as a double precision floating point number or with the precision of the result item.