A computer chip that performs floating-point division, where decimal points are removed before calculation and inserted after computation. In computer graphics, the use of a math coprocessor speeds up calculations when performing a number of functions, cutting response times by up to 90 percent. IBM-compatible computers with 80486DX or Pentium microprocessors have a coprocessor included in the main processing chip. So do Macintosh computers with PowerPC microprocessors; standard Macintosh computers and IBM-types with SX or 80386 (or older) microprocessors need separate math co-processing chips. Also called a floating-point unit (FPU) chip with Macintoshes.
A chip that relieves the computer's microprocessor of numeric-processing tasks. The IntelĀ® PentiumĀ® III and Celeron(tm) microprocessors, for example, include a built-in math coprocessor.
A math coprocessor (aka: FPU, NPU, math-co) is a chip designed to perform floating point calculations very quickly. Sometimes the main processor does not include an onboard FPU in which case it executes floating point instructions very slowly compared to a FPU. 486 computers have the co-processor built in, while it must be added to a 386.
A computer chip designed especially to process numeric-intensive instructions faster than ordinary chips; used in conjunction with standard chips.