double precision

views updated

double precision The use of double the usual number of bits to represent a number. Arithmetic performed on double-precision numbers is called double-precision (or double-length) arithmetic. For floating-point numbers, most computers use the same number of bits for the exponent in single-length and double-length forms. Consequently, if the length of a single-precision number is l bits, p of which are used for the significand, then the significand of a double-precision number occupies (p + l) of the 2l bits. Occasionally, multiple precision, i.e. more than double precision, may be available. Some computers implement double precision in hardware; higher precision, for example quadruple precision, is almost always achieved by software.