【摘要】 采用硬晶片的三维堆叠SoC测试规划是一个NP hard问题,针对该问题提出了一种采用GWO(grey wolf optimization)的三维堆叠SoC测试规划方法,使得在最大测试引脚数和最大可使用TSV(through silicon vias)数的约束条件下,从而达到三维堆叠SoC测试时间最小化目的。本算法基于群体智能,通过实施攻击等操作,更新Alpha、Beta和Delta进行寻优,从而实现三维堆叠SoC测试规划。本研究以ITC’02 Test benchmarks中的典型SoC为实验堆叠对象,实验结果表明本算法相比PSO(particle swarm optimization),能够获得更短的测试时间。更多还原

【Abstract】 Test scheduling of three-dimensional stacked SoC is a NP hard problem. This paper proposes a GWO( grey wolf optimization) algorithm for the three-dimensional test scheduling,which can achieve the purpose of minimization of test time of the three-dimensional integrated circuit under the condition of the maximum number of TSV available and the maximum number of test pins available. Through the implementation of attack operation,the algorithm based on swarm intelligence can achieve the test scheduling of three-dimensional stacked SoC. Typical SoC in ITC’02 benchmark is adopted,and the experimental results show that the algorithm can obtain shorter test time compared with PSO( particle swarm optimization).更多还原