IEEE Floating Point Standard

IEEE Floating Point Standard

(standard, mathematics)(IEEE 754) "IEEE Standard for BinaryFloating-Point Arithmetic (ANSI/IEEE Std 754-1985)" or IEC559: "Binary floating-point arithmetic for microprocessorsystems". A standard, used by many CPUs and FPUs, whichdefines formats for representing floating-point numbers;representations of special values (e.g. infinity, very smallvalues, NaN); five exceptions, when they occur, and whathappens when they do occur; four rounding modes; and a setof floating-point operations that will work identically on anyconforming system.

IEEE 754 specifies formats for representing floating-pointvalues: single-precision (32-bit) is required,double-precision (64-bit) is optional. The standard alsomentions that some implementations may include single-extendedprecision (80-bit) and double-extended precision (128-bit)formats.