LINEAR DEPENDENCY CHECK IN BUILT-IN TEST

Abstract

Given a characteristic polynomial p(x) and a sampling polynomial with respect to p(x), a designer likes to know whether the sampling polynomial is a dependency polynomial. A dependency polynomial gives low fault coverage in the built-in self test. The tool LDC computes the residue combinations of the factors in the set polynomial and advises designers if the sampling polynomial covers all the testing space. When the sampling polynomial is a dependency polynomial, LDC will try to find one, if any, independent polynomial at a minimum cost. In cases of short test vectors LDC can exhaustively enumerate all independent polynomials for the designers to choose from. This paper reviews the basics of the linear dependency computation, explains the LDC capabilities and its usefulness, and discusses the computational issues of the problem.