论文部分内容阅读
分组密码IP核具有配置过程复杂、数据运算量大的特点,如何对其进行高效的验证是整个设计面临的关键问题.在随机验证中,激励生成和覆盖率模型抽象占据着尤为重要的位置.本文将分类树方法应用于分组密码IP的功能验证,并且针对其无法解决关联操作和顺序控制的缺点实施改进,主要是引入虚拟输入对激励序列进行规划,构建超长输入数据包.实验证明,采用改进的分类树指导激励生成和覆盖率模型抽象,能够生成更加精简有效的激励和完备的覆盖率模型,进而显著地提高验证的效率和完备性.
Packet cipher IP core has the characteristics of complicated configuration process and large amount of data computation, and how to verify it efficiently is a key issue facing the whole design.In the random verification, incentive generation and coverage model abstraction occupy a particularly important position. In this paper, the classification tree method is applied to functional verification of block cipher IP, and the improvement of its shortcomings that it can not solve the associated operation and sequence control is mainly to introduce the virtual input to plan the excitation sequence and construct long input data packet.Experiments show that, Adopting improved classification tree to guide the generation of incentive and abstraction of abstraction model can generate a more streamlined and effective motivation and a complete coverage model, thereby significantly improving the efficiency and completeness of verification.