论文部分内容阅读
针对片上网络(NoC)中大量节点的测试难题,提出了一种结合二维云进化算法优化选取NoC中测试端口位置,提高测试效率的方法。该方法结合NoC网格结构特点,采用重用测试访问机制和XY路由方式,由测试功耗限制确定端口对数,通过二维云模型对端口坐标进行统一建模,云进化算法自适应控制遗传变异的程度和搜索空间的范围,在测试功耗约束条件下,优化选取最佳测试端口的位置,达到总测试时间最少的目的。以SoCIN结构电路为仿真平台,分别对4×4网格和8×8网格结构NoC进行了实验仿真,结果表明,在NoC节点测试问题上,云进化算法能快速收敛到最优解,有效提高整体测试效率。
Aiming at the problem of testing a large number of nodes in NoC, this paper proposes a method combining two-dimensional cloud evolution algorithm to optimize the location of test ports in NoC and improve the test efficiency. According to the characteristics of NoC grid structure, this method adopts the reusable test access mechanism and the XY routing method, the logarithm of the port is determined by the test power consumption limit, and the port coordinates are uniformly modeled by a two-dimensional cloud model. The cloud evolution algorithm adaptively controls the genetic variation Of the degree and the scope of the search space, optimize the location of the test port under the test power constraints, to achieve the purpose of the minimum total test time. Taking SoCIN structure circuit as the simulation platform, the simulation experiments of 4 × 4 grid and 8 × 8 grid NoC were carried out respectively. The results show that the cloud evolution algorithm converges to the optimal solution quickly and effectively Improve the overall test efficiency.