摘要
提出了基于小生境遗传算法的系统级芯片(SoC)测试存取机制(TAM)的优化方法.结合TAM宽度约束进行SoC中功能内核(IP)的测试壳的优化,解决测试存取机制的测试总线划分及测试总线细分等的优化问题,取得了较好的结果,并有效地减少SoC的测试时间.采用分支-联合(Fork-Joint)的方法可得到更为优化的TAM方法,对于ITC 2002基准SoC d695,比未采用分支-联合方法的TAM划分方法的测试性能最大可以提高30%,和其它方法的优化结果相比,该方法的平均效果优于其它方法1到9个百分点.
Optimization schemes based, on niche genetic algorithm were presented in this paper to solve TAM co-optimization issues such as TAM partition and subdivision when wrapper optimization issue under constraints of width of TAM was considered. These schemes had better results and minimized the testing time for a SoC effectively. When the fork-joint scheme was used, better results were achieved about 30% promotion compared to schemes without fork-joint for ITC 2002 benchmark SoC d695. The scheme in this paper had 1% - 9% promotion than other schemes reported currently.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2007年第5期825-829,共5页
Journal of Harbin Institute of Technology