MINT--An Exact Algorithm for Finding Minimum Test Set--


IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E76-A   No.10   pp.1652-1658
Publication Date: 1993/10/25
Online ISSN: 
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on VLSI Design and CAD Algorithms)
test pattern generation,  minimum test set,  binary decision diagram,  minimum set covering problem,  

Full Text: PDF(561.6KB)>>
Buy this Article

In this paper, an exact algorithm for finding minimum test set which detects all testable stuck-at faults of a given combinational circit is presented. So far several heuristic algorithms for this problem are proposed, but no efficient exact algorithms are known. To solve this exactly, minimum test set problem is formalized as a minimum set covering problem, and then implicit manipulation technique using binary decision diagrams(BDDs) is applied. The algorithm presented in the paper has two contributions. One is utilization of maximal compatible fault set, which can drastically reduce the number of candidates for minimum test set. A new BDD based algorithm for extracting all maximal compatible fault sets is shown. The other is a new implicit manipulation technique handling with huge covering matrix. Actually, the algorithm using this technique can handle minimum set covrering problem with over ten thousand columns in a few minutes. Experiments using ISCAS benchmark circuits show that the algorithm is quite efficient for small(100-300 gates) circuits. A computational complexith of minimum test set problen is much higher than that of ordinary test pattern generation problem, so that practical signifcance of this method is not high. But the algorithm is still useful for evaluation of other heuristic algorithms. furthermore, this implicit manipulation technique can also be applied to other minimumset covering problems.