Efficient Class of Redundant Residue Number System
Somayeh Timarchi, Keivan Navi · 2007
The residue number system is suitable for implementing high-speed digital processing devices because it supports parallel, modular, fault-tolerant and carry-bounded arithmetic. The carry propagation is restricted to inside the modulus. The remaining intra moduli carry propagation limits the speed of arithmetic operation. Therefore the carry-free property of redundant arithmetic can be used. In this paper, we discuss a novel class of redundant high-radix RNS based on the stored-unibit transfer representation for modulo 2n+1 that improves the performance of conventional redundant RNS and provides a desired tradeoff between area and delay complexity.