Design Of Modulo Multipliers And Residue-to-binary Converter On Residue Number System

Due to its inherent parallelism, free carry properties, RNS has potentially been widelyusing in the field of digital signal processing which has intensive addition and multiplicationoperations Our work is carry out in researching modulo set selection for the specificapplication, and modulo multipliers, residue to binary converterFirst＄After analyzing the performance of FIR based on RNS, we consider the64th orderdual bandpass linear phase filter as ours specific application, we make an comparsion frommulti dimension (include performance of modulo operation unit, the balance betweenmodulos, etc) among many modulo sets We select modulo set {2n+11,2n,2n1} as ourresearch modulo set in the background of dynamic rang width equal to19Second, we combine radix4Booth codec method and method of scaling partial productmatrix to propose a new module2k1multiplier There are some ‘0’ constant occur in thescale down partial product matrix, so we can further reduce area cost when we make Boothcodec compare with previous modulo2k1multiplier Theoretical analysis indicates when ‘k’is odd, the performance of our new module2k1multiplier is better than other previouspublished, and experimental results show our new module2k1multiplier can respectivelymake89%,1074%area reduction, without any delay penaltyThird, we propose an area time efficient RNS reverse converters for modulo set {2n+11,2n,2n1} which is profitable to construct an efficient residue number system for free of2n+1modulus Based on the modified Chinese Remainder Theorem (CRT) and with a differentorder (we treat2n+11first), we only need one mod arithmetic to implement our reverseconverter and we further reduce modulo size that only requires arithmetic mod (2n1)operation Our new reverse converter is completely composed of adders and few inverterswithout any other logic circuits Theoretical analysis and experimental results assert that ournew reverse converter costs less hardware area, and can provide4175%＄6025%＄625%＄5475%＄and24%area*time reduction respectively compare with previous published...