Improving Goldschmidt division, square root, and square root reciprocal

The aim of this paper is to accelerate division, square root, and square ro ot reciprocal computations when the Goldschmidt method is used on a pipelin ed multiplier. This is done by replacing the last iteration by the addition of a correcting term that can be looked up during the early iterations. We describe several variants of the Goldschmidt algorithm, assuming 4-cycle p ipelined multiplier. and discuss obtained number of cycles and error achiev ed. Extensions to other than 4-cycle multipliers are given. If we call G(m) the Goldschmidt algorithm with m iterations, our variants allow us to reac h an accuracy that is between that of G(3) and that of G(4), with a number of cycle equal to that of G(3).