基于粒子群算法的多约束3D NoC协同测试规划

发布时间：2018-01-11 03:07

本文关键词：基于粒子群算法的多约束3D NoC协同测试规划　出处：《仪器仪表学报》2017年03期 　论文类型：期刊论文

【摘要】：为了提高三维片上网络(3D NoC)资源内核的测试效率,对多约束下的3D NoC进行测试规划。在硅通孔(TSV)数量、功耗以及带宽约束下,分别将TSV位置、IP核测试数据分配作为两个寻优变量,利用离散粒子群算法协同进化,以减少测试时间并提高TSV利用率。在算法中引入全局次优极值对粒子进行指导,提高全局搜索能力;并通过自适应参数调整策略增加种群多样性,从而改善粒子搜索的停滞现象。以国际标准测试集ITC'02中的电路作为仿真对象,仿真结果表明,算法能够有效地完成在多约束下对TSV位置的寻优并合理分配通信资源,缩短了测试时间,提高了TSV利用率。
[Abstract]:In order to improve the testing efficiency of 3D NOC resource kernel, the test planning of 3D NoC with multiple constraints is carried out under the constraints of the number, power consumption and bandwidth of the via hole. Using TSV location IP core test data allocation as two optimization variables, discrete particle swarm optimization algorithm is used to co-evolve. In order to reduce the test time and improve the utilization of TSV, the global suboptimal extremum is introduced to guide the particle to improve the global search ability. In order to improve the stagnation of particle search, the adaptive parameter adjustment strategy is used to increase population diversity. The simulation results show that the circuit in the international standard test set ITC'02 is used as the simulation object. The algorithm can effectively optimize the location of TSV under multiple constraints, allocate communication resources reasonably, shorten the test time and improve the utilization of TSV.
【作者单位】：桂林电子科技大学电子工程与自动化学院;广西自动检测技术与仪器重点实验室;
【基金】：国家自然科学基金(61561012)项目资助
【分类号】：TN47;TP18
【正文快照】： 1引言随着三维集成电路技术的不断发展,三维片上网络(three dimensional network-on-chip,3D No C)应运而生。它通过硅通孔(through-silicon-via,TSV)技术实现各层二维结构芯片的互连,具有互连线短、集成度高、功耗低、扩展性好等优势[1-2]。但是,一方面,由于TSV造价高昂、占

【相似文献】