Effective implementation of Wang method using approximate method in RNS

Abstract

In the paper we propose an effective implementation of Wang method using approximate method in Residue Number System (RNS). Approximate method allows to substitute the computationally complex operation of finding residue of division by a large number with taking lower bits. We show that for RNS to binary conversion with modified Wang method the sufficient precision of computation of coefficients is [log2 M]. Use of Wang method together with approximate method allows to decrease the size of numbers from [log2 (μ · M)] to [log2 M]