Area-Efficient Multi-moduli Squarers for RNS

Abstract

Multi-moduli architectures are very useful for reconfigurable digital processors and fault-tolerant systems that are based on the Residue Number System (RNS). In this paper we propose two architectures for multi-moduli squaring that support the most common moduli cases in RNS channels, that is, 2^n-1, 2^n and 2^n+1. The proposed architectures are based on the modified Booth encoding of the input operand for deriving the required partial products and on Dadda adder trees for their addition. Experimental results show that the proposed squarers offer significant savings in area compared to previous proposals while a small improvement in delay is achieved in most cases as well.