US20020116431

System and method for improving the accuracy of reciprocal operations performed by a floating-point unit