An additional microprocessor installed in a computer so as to supplement the central processing unit and expedite and facilitate certain mathematical functions (often by being designed to perform exponential arithmetic, expediting the calculation of large numbers). When the CPU encounters an instruction intended for a coprocessor, it sends it to the coprocessor and awaits a signal indicating that the calculation is completed. (If there is no coprocessor installed, an error message may be the result.) Since these coprocessors are designed specifically for certain types of calculations, they require significantly fewer instructions than a general CPU would. Many newer CPUs are now created to perform these functions. Spreadsheets, CAD, and some graphics programs perform significantly better when coprocessors are installed, although a program must be written specifically to take advantage of a coprocessor.
These coprocessors are sometimes called floating-point coprocessors, or floating-point units.