System for testing of digital integrated circuits

摘要

In integrated circuit testing which involves partitioning of tailored test vectors into subsets and filling of the subsets with similar vectors starting with vectors close to the initially selected vector and continuing to add vectors farther away until an optimal number of vectors are in the subset, a system is provided for selecting a number of test vectors to go into a subset by using a unique "distance" measure to measure how far away the candidate vector is from what is already in the subset. Distance determination involves comparing the candidate vector with the weights generated from randomization of the previous vectors in the subset. Having provided a unique distance or "closeness" measure for each candidate vector, the distance of a number of candidate vectors is ascertained. The one with the smallest distance from those vectors already in the subset is selected for determination as to whether its inclusion in the subset results in more rapid and accurate fault detection. This determination is made based on a prediction algorithm which includes calculation of the mathematical expectation of the number of tailored vectors not appearing in a weighted random sequence.