摘要
采用带分复用的测试方法,结合自然进化策略的云进化算法(CEA)优化3D NoC测试规划,依据带分复用的特点建立3D NoC测试时间优化模型,在功耗和带宽双重约束下,采用改进的云进化算法(ICEA)对该模型进行求解,完成3D NoC测试.以ITC’02基准电路为实验对象,实验结果表明该算法能迅速收敛到适应值最佳的测试规划,实现3D NoC资源内核的最大化并行测试,提高测试效率,缩短测试时间.
This paper adopt the test method of bandwidth division multiplexing,based on cloud evolutionary algorithm that combined with the natural evolution strategies to optimize the test scheduling for 3DNoC.Establish test time model according to the features of bandwidth division multiplexing,under the dual constraints of power and bandwidth,the model is solved with improved cloud evolutionary algorithm to complete the test.With ITC＇02test circuit as the simulation object,experimental results show this algorithm can search best test scheduling quickly,maximize the parallel test on 3D NoC resources cores,improve test efficiency and reduce testing time.
出处
《微电子学与计算机》
CSCD
北大核心
2016年第7期74-78,共5页
Microelectronics & Computer
关键词
带分复用
双重约束
改进云进化算法
测试规划
bandwidth division multiplexing
dual constraints
improved cloud evolutionary algorithm
test scheduling
