An improved residue multiplier using Montgomery modular multiplication

Implementing cryptosystems using the residue number system (RNS) requires many balanced equally-wide residue channels. The generic residue multiplier has a significant impact on the residue balance. Improving the residue multiplier delay affects the RNS efficiency. The Montgomery reduction is used for the residue gen-eration in the auto-scale multiplier method, where its operations are merged to reduce residue generation delay. The synthesis result shows the proposed work delay has improved by 36%, 41%, and 29% for three channel widths of 16, 32, and 64-bit compared to the previous relevant work. Moreover, the area consumption has reduced by about 40%.