(computer science) Any calculation that is performed on numbers whose digits of precision are limited only by the available memory of the host system.