许川佩, 陈玄. 基于猴群算法的3D NoC IP核测试优化方法[J]. 微电子学与计算机, 2019, 36(1): 22-26, 31.
引用本文: 许川佩, 陈玄. 基于猴群算法的3D NoC IP核测试优化方法[J]. 微电子学与计算机, 2019, 36(1): 22-26, 31.
XU Chuan-pei, CHEN Xuan. IP Cores Test Optimization Method of 3D NoC Based on Monkey Algorithm[J]. Microelectronics & Computer, 2019, 36(1): 22-26, 31.
Citation: XU Chuan-pei, CHEN Xuan. IP Cores Test Optimization Method of 3D NoC Based on Monkey Algorithm[J]. Microelectronics & Computer, 2019, 36(1): 22-26, 31.

基于猴群算法的3D NoC IP核测试优化方法

IP Cores Test Optimization Method of 3D NoC Based on Monkey Algorithm

  • 摘要: 如何对三维片上网络(three Dimensional Network-on-Chip, 3D NoC)资源内核的测试进行优化以缩短测试时间, 提高资源利用率是当前3D NoC测试面临的主要问题之一.本文针对3D NoC IP核测试优化问题, 开展TSV位置与IP核测试数据分配方案协同优化研究.在带宽、功耗和TSV数量约束下, 将TSV位置方案和IP核测试数据分配方案作为寻优变量, 采用猴群算法进行寻优.算法通过爬和望跳过程进行局部搜索并结合翻过程在不同领域进行搜索从而找到最优解, 加入精英保留策略以确保算法收敛性, 使算法搜索结果更为准确.以ITC’02电路为实验对象, 实验结果表明, 该算法能够有效地优化3D NoC资源分配, 缩短测试时间, 提高资源利用率.

     

    Abstract: How to optimize the test of the three-dimensional network-on-chip (3D NoC) resource core to shorten test time and increase resource utilization is one of the major problems that the 3D NoC testing faced. This work focuses on 3D NoC IP core test optimization issues, and conducts collaborative optimization research on TSV location and IP core test data distribution schemes. Under the constraints of bandwidth, power consumption and TSV quantity, the TSV location scheme and IP core test data distribution scheme are used as optimization variables and Monkey Algorithm(MA) is used to optimize. The algorithm performs local search through the climb and watch-jump process and searches in different fields with the somersault process to find the optimal solution. The elite retention strategy is added to ensure the convergence of the algorithm and the algorithm search result is more accurate. Taking the ITC'02 circuit as the experimental object, the experimental results show that the algorithm can effectively optimize the 3D NoC resource allocation, shorten the test time, and improve resource utilization.

     

/

返回文章
返回