Effective RNS scaling algorithm with the Chinese remainder theorem decomposition

A novel scaling technique in the residue number system (RNS) is proposed. In this technique, the main computational effort is made in the precomputing phase. The remainder calculations are performed by the modulo and binary adders. The scaling requires one look-up cycle and time for modulo addition of the n+2 operands. In the proposed approach there are restrictions neither on the form and the size nor on the number of moduli of the RNS. The scaling factor K can be integer or real, and it must fulfill merely a weak condition K > n, where n is the number of moduli. The absolute scaling error by using the correction scheme does not exceed 1.5.<
>