TABLE OF CONTENTS A carry-skip adder reduces the time needed to propagate the carry by skipping over groups of consecutive adder stages. As such, the carry-skip adder generalizes the idea behind the Manchester adder described in Section 5.4. The carryskip adder illustrates the dependence of the optimal algorithm for addition on the available technology. Although the carry-skip algorithm has been known for many years, it has become popular only recently. In VLSI technology the carry-skip adder is comparable in speed to the carry look-ahead technique (for commonly used word lengths but not necessarily in the asymptotic sense) but it requires less chip area and consumes less power.
The carry-skip adder is based on the following observation. The carry propagation process can skip any adder stage for which x m ?y m (or in other words, P m =x m ? y m=1). Several consecutive stages can be skipped if all satisfy x m ?ym. Thus, an adder consisting of n stages is divided into groups of consecutive stages with a simple ripple-carry scheme used in each group. Every group also generates a group-carry-propagate signal that equals 1 if all stages internal to the group satisfy P m=1. This signal can be used to allow an incoming carry into the group to skip all the stages within the group and generate a group-carry-out. Let a particular group, say, group l, consist of the k bit positions j, j+1, , j+
