(computing) A number storage format that uses two adjacent locations in memory, allowing for greater precision in the values stored.