可重构分组密码逻辑阵列加权度量模型及高能效映射算法

针对基于粗粒度可重构阵列结构的分组密码算法映射情况复杂、难以实现统一度量的问题，该文采用多目标决策手段，以性能及功耗参数为决策目标，基于分组密码算法轮运算及粗粒度可重构阵列结构特征约束，提出了一种面向分组密码算法映射的加权度量模型.同时，采用主客观综合分析法，定义了模型权重参数的计算方式，从而通过配置合理的权重参数，以高能效映射算法实现差异化的映射.为了降低决策时间，该文进一步提出了基于二进制编码的枚举搜索算法，实现了最优映射结果搜索与映射矩阵建立的并行，使决策的时间复杂度降至O（2ⁿ）.实验结果表明，该文提出的加权度量模型能实现高效的分组密码算法映射方案决策，单位面积性能提升了约14.2%，能效提升了约一倍.

Reconfigurable Block Cryptographic Logic Array Weighted Metric Model and High Energy-Efficient Mapping Algorithm

Aiming at the problem that the block cipher algorithm mapping based on coarse-grained reconfigurable array structure is complex and the evaluation standard is not uniform,this paper adopted a multi-objective decision-making means to evaluate the performance and power consumption parameters,and proposed a weighted metric model for map of block cipher algorithm.At the same time,the method of weighting parameters is defined by comprehensive subjective and objective factors,so as to provide differentiated mapping scheme by configuring reasonable weight parameters.In order to reduce the decision time,this paper introduced the algorithm of enumeration search based on binary coding,and realized the parallelism between the optimal scheme search and the generate mapping matrix,so that the time complexity of decision decrease to O(2 ⁿ).The experimental results show that the weighted metric model can achieve efficient block cipher algorithm mapping,and the throughout per chip area has improved about 14.2%,with a 100% improvement in energy efficiency per workload bit.