A floating-point unit (FPU, colloquially a math co-processor) is a part of a computer system specially designed to carry out operations on floating-point numbers. Typical operations are addition, subtraction, multiplication, division, and square root. Some FPUs can also perform various transcendental functions such as exponential or trigonometric calculations, but the accuracy can be very low, so that some systems prefer to compute these functions in software.
In general-purpose computer architectures, one or more FPUs may be integrated as execution units within the central processing unit; however, many embedded processors do not have hardware support for floating-point operations (while they increasingly have them as standard, at least 32-bit ones).
When a CPU is executing a program that calls for a floating-point operation, there are three ways to carry it out:
- A floating-point unit emulator (or floating-point library), such as SANE, SoftwareFPU, or PowerFPU.
- An add-on FPU, a dedicated hardware unit such as the Motorola 68881 or 68882.
- An integrated FPU, included in processors such as the Motorola 68040, and all PowerPC and Intel processors used by Apple, which made separate add-on FPUs obsolete.
- "The IBM System/360 Model 91: Floating-Point Execution Unit", IBM Journal of Research and Development, January 1967, pp. 34–53.
- Bruce Dawson (2014-10-09). Intel Underestimates Error Bounds by 1.3 quintillion. randomascii.wordpress.com. Retrieved on 2020-01-16.
- FSIN Documentation Improvements in the “Intel® 64 and IA-32 Architectures Software Developer’s Manual”. intel.com (2014-10-09). Retrieved on 2020-01-16.