Simple accuracy
In mathematics and computer science is single precision ( English single precision or only single ) is a name for a floating point format , which has a storage unit in the computer. The exact details therefore depend on the computer manufacturer and its memory architecture. The IEEE 754 standard was specially designed for microprocessors with byte-by-byte memory, which stipulates 4 bytes (32 bits) for this number format. The designation is not reserved for floating point numbers; it can also be used for whole number formats.
An IEEE-754 number has a precision between 7 and 8 valid digits in the decimal system .
Often there is a need to calculate results with higher precision, which is why there is the double precision number format .
Since programming languages such as C treat single precision as a matter of principle, the double memory consumption only plays a minor role today and the higher computing time is hardly of any consequence, the same numerical problem arises for double precision as for simple. Therefore, in the revision IEEE 754-2008, four-fold exact number formats were introduced.
For special numerical tasks, e.g. B. in computer graphics and for didactic purposes, there are also shorter number formats than simple precision.