Low-power modified shift-add multiplier design using parallel prefix adder

Multipliers are the building blocks of every digital signal processor (DSP). The performance of any digital system is dependent on the adder design and to a large extent on the multiplier block. Area and power dissipation are the major considerations in a multiplier design due to its complexity. In this paper, an efficient multiplier "bypass zero feed multiplicand directly," based on shift-add multiplication, has been proposed for low-power application. As the shift-add multiplication is a repetitive process of addition, the implementation time of an adder is reduced by using the proposed parallel prefix adders designed based on revised Ling equations. The proposed 8-bit, 16-bit and 32-bit multipliers are implemented using 180-nm and 90-nm CMOS technologies. Simulation results reveal that the proposed multiplier is fast and lowers the power by 35% predominantly for a 32-bit multiplier.