摘要
在SoC的测试规划时,考虑为避免在测试过程中出现热点以及测试过程中使热量均匀分布,基于建立的问题模型得到一系列的并行测试集合,再通过Bin-Packing算法构造测试规划,并进行全局的优化.对ITC’02测试用例的实验结果表明,该方法在牺牲一定的测试时间的情况下,有效地控制了在测试时芯片温度的升高,从而避免出现由热量引起的一系列问题.
An approach based on Bin-Packing algorithm and graph theory is proposed to avoid the hotspot and distribute heat evenly on chip surface during test scheduling. First several parallel test set (PTSs) are derived from graph models. Then Bin-Packing algorithm constructs an initial test schedule. A global optimal procedure gets the better test schedule finally. Experiment at results on ITC'02 benchmark SoCs show that this method can effectively avoid hot-spot and distribute heat evenly at the expense of a little percentage of test time.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2006年第1期46-52,共7页
Journal of Computer-Aided Design & Computer Graphics
基金
国家"八六三"高技术研究发展计划(2003AA1Z1120
2004AA1Z1050)
国家自然科学基金(90307017
60176017
90207002)
美国国家科学基金(NSFgrantsCCR-0098275
CCR-0306298)