一种基于遗传算法的多IP核并行测试方法

A Multiple IP Core Concurrent Test Method Based on Genetic Algorithm

针对现有SoC测试方法所需测试时间过长的问题,提出了一种基于遗传算法的多IP核并行测试方法.该方法主要是在功耗约束的情况下,通过遗传算法将尽可能多的IP核的测试数据压缩,即IP核的测试数据相同的位接到同一根总线数据位,使每次并行测试的IP核数量达到最大.此方法不仅可以减少测试数据集的大小,且能够减少对总线位宽的需求.通过应用遗传算法对所建立测试数据求取最佳测试方案的仿真,证明了该方法是可行的,且能够极大缩短测试时间.

Aiming at the problem of the existing SoC test methods required long testing time,this paper proposes a multiple IP core concurrent test method based on genetic algorithm.This method is mainly under the condition of the power consumption constraints,through the genetic algorithm as much as possible of the IP core test data compression.In other words,a bus access multiple IP core with the same test data,make each time the number of IP core of concurrent test to achieve the most.This method not only can reduce the size of the test data set,and can reduce the demand for bus bits wide.By using genetic algorithm to calculate the best test scheme of simulation show that the method is feasible,and can greatly shorten the test time.