分析了Montgomery模乘算法及其几种实现方式,指出FIPS方式是适合乘加器结构VLSI实现的一种算法.给出了FIPS方式的数据通路和控制部分的实现方案.提出了在选择不同的操作数位宽的情况下,对具体实现的评价标准.结合具体数据分析了随着操作数位宽的变化,面积、速度和功耗指标的变化趋势,并对使用单乘法器和双乘法器的情况进行了比较.
In this article, we analyze Montgomery multiplication and some ways to perform it. It is pointed out that FIPS mode is suitable for the implementation based on VLSI multiplier and accumulator structure. The data path and control structure based on FIPS mode are designed. We discuss the evaluation standard for the implementations based on different operand width. By using the experimental data, we analyze the differences in size, speed and power of the implementations based on different operand width. We also compare single multiplier mode with double multiplier mode.