面向总线系统的高层次结构化激励生成算法

为了应对大规模设计中逻辑信号级输入激励空间爆炸的问题,针对总线系统提出了一种高层次结构化激励生成算法和相应的功能覆盖率模型.首先将总线系统抽象成通用有向二分图模型,然后建立相应激励的高层次数学模型,由此提出一种通用的层次化输入激励空间等价类划分算法和对应的高层次功能覆盖率模型,最后基于树的搜索提出了2种结构化激励生成算法.上述方案成功应用于IME-Diamond SoC的总线系统的功能验证中,实际结果表明,相比代码覆盖率,高层次功能覆盖率模型的揭示功能Bug能力更强,而且相对于传统的随机生成,结构化的激励生成能够将覆盖率收敛所需的激励数减少96%.

Bus System Oriented High Level Structural Stimuli Generation Algorithm

In order to deal with the problem of the explosion of the signal level input stimuli space in the massive design, a high level structural stimuli generation algorithm and corresponding functional coverage model targeted at bus system are presented. Firstly, the bus system is abstracted to be a general bipartite graph model (BGM), and then, the mathematical model of the stimuli is established. Consequently, a general layered equivalence partition algorithm of ISS and the relevant high level functional coverage model are proposed. Finally, two structural stimuli generation algorithms based on the search of ISS tree are presented. Experiments are performed in the functional verification of bus system of IME-Diamond SoC. The result indicates that the high level functional coverage model can help uncover functional bugs more easily when compared to code coverage, and structural stimuli generation can reduce 96% stimuli for coverage convergence when compared to random stimuli generation.