Realization of intelligent optimization algorithm on IP cores partition for NoC testing
Ling Yunhui, Fang Liu, Ying Zhang · 2015
According to the principles of nature or the biosphere, people imitate its rules to create the intelligent algorithm for solving problems. Nowadays there are simulated annealing algorithm, genetic algorithms, artificial neural networks, swarm intelligence algorithms and so on. With the specific testing access architecture, testing embedded IP cores of NoC-based SoC can be altered to IP cores partition problem. The optimization object is to minimize the parallel testing time. The paper proposed the realization of intelligent optimization algorithm to group IP cores for NoC testing by artificial neural network method. The artificial neural network is an algorithm mathematical model of distributed parallel information processing. The experiment results of ITC'02 benchmark circuits on MATLAB showed the effectiveness of the proposed intelligent optimization algorithm. The parallel testing time of embedded IP cores on NoCs decrease by averagely 5.857% (W=16) and 5.65% (W=32).