On the impact of encoding on the complexity of residue arithmetic circuits

In this paper we explore the impact of data encoding on the design of arithmetic circuits based on residue number system (RNS). Specifically, we show that departing from the conventional representation of residues as integers in weighed binary format and appropriately selecting a non-weighed encoding, efficient circuits in terms of power, area and performance are possible. A quantitative analysis reports savings of the order of 45% for some complexity-related metrics of practical interest.