Correlation Analysis of Compacted Test Vectors and the Use of Correlated Vectors for Test Generation

In this paper we present novel correlation analyses methods to analyze the compacted test sets. In spatial correlation, we identified strong correlation chains, while in the temporal domain, correlating a segment of the test set with the entire sequence discovered periodicities in the compacted test set sequence. These suggest the potential of a new way for generating test sequences. We report preliminary experiments in which only the spatial correlation was used to generate random vectors, which when compacted, provided tests for several ISCAS'89 benchmarks.