Twice the accuracy

Format of a number in double precision according to IEEE 754

Double precision ( English double precision or double ) is on the computer arithmetic for a floating point -Zahlenformat in which a number of 8 bytes (i.e. 64 bits ) occupied. It therefore occupies twice as much memory as single-precision floating point numbers . In the floating point representation, 11 bits are used for the exponent and one bit for the sign, the remaining 52 bits are available for the actual number representation ( mantissa ). A precision of 53 bits, converted into the decimal system, means an approximate accuracy of 16 places in the decimal system ( ). ${\displaystyle 53\log _{10}(2)\approx 15.955}$

The designation is not only reserved for floating point numbers, but can also be used for whole number formats.

The representation of such floating point numbers has been specified in the 754 standard by the Institute of Electrical and Electronics Engineers since 1985  .

Further accuracy classes

• Half accuracy or mini floats
• Simple accuracy
• Quadruple accuracy

Individual evidence

1. ↑ 754-1985 - IEEE Standard for Binary Floating-Point Arithmetic. Retrieved January 15, 2018 .