Chip Test Pattern Reordering Method using Adaptive Test to Reduce Cost for Testing of ICs

Abstract

With the continuous drive toward integrated circuits (ICs) scaling, more test patterns are required in testing. However, the large number of patterns continues to increase test time and test costs. Thus, test costs of ICs are becoming more crucial yet more challenging. In this paper, we propose a novel adaptive test strategy to reduce test costs without increasing test escape, and using shortest path first (SPF) algorithm combined with K-Nearest Neighbor (KNN) to reorder the test patterns. The patterns which identify faults earlier are moved forward and applied first so as to save test time. In addition, the optimal test patterns are searched by means of a polynomial regression function and it provides a trade-off between test cost and test escape. Consequently, the optimization problem is converted into a mathematical function, in order to achieve broader applicability and generality. Experimental results demonstrate that the proposed method achieved 45.2 seconds time savings but leads to almost negligible test escapes increasing compared with traditional methods. The test pattern reordering method aims to find defects as early as possible. Furthermore, the proposed algorithm is completely software based and incurs no additional hardware overhead.