A floating point unit (FPU), also known as a math coprocessor or numeric coprocessor, is a specialized coprocessor that manipulates numbers more quickly than the basic microprocessor circuitry. The FPU does this by means of instructions that focus entirely on large mathematical operations. In the early years of personal computing, the FPU was physically separate from the main microprocessor. Starting with the Intel Pentium and Motorola 68000 series in the late 1990s, the FPU became a physical part of the microprocessor chip.
A floating point number is a binary number carried out to a specific number of digits, and containing a radix (or decimal) point somewhere in the digit sequence. Such numbers are stored in three parts: the sign (either plus or minus), the mantissa (sequence of meaningful digits), and the exponent (power or order of magnititude), which determines the position in the mantissa where the radix point is placed. The main operations of the FPU consist of conventional arithmetic such as addition and multiplication. Some FPUs can also perform more sophisticated functions such as exponentials, logarithms, and trigonometry.
Some programs commonly downloaded from the Internet, such as Macromedia's Shockwave, require that a computer have an FPU. If a computer is old and its microprocessor does not have an FPU built in, a program called an FPU emulator can be downloaded. This can, in many cases, allow such programs to run, but not as fast as they would in a computer that has a hardware FPU.