Design and analysis of a low-cost approximate adder with OR and zero truncation

This paper proposes a new cost-effective approximate adder that exploits OR operation and zero truncation. The proposed approximation technique reduces the hardware cost significantly while maintaining comparable computation accuracy. The proposed adder achieved 48%, 51%, and 48% reductions in the area, delay, and power, respectively, compared to a traditional adder when implemented in 32-nm CMOS technology. The proposed design could also enhance the normalized mean error distance up to 29% compare to the approximate adders considered in this paper. The adder showed an excellent tradeoff performance between the hardware and computation accuracy. Furthermore, the proposed adder was adopted in a digital image processing application, and the benefit of the proposed adder is demonstrated.