SEARCH STATE EQUIVALENCE FOR REDUNDANCY IDENTIFICATION AND TEST GENERATION

Abstract

We present new extensions to the EST' algoritlm, which accelerates combinational circuit Redundancy Identification and Automatic Test Pattern Generation (ATPG) algorithms, in particular SOCRATES. EST detects equivalent search states, which are saved. for all faults during ATPG. The search space is reduced by using learned Search State equiualences to detect previously-encountered search states (possibly from prior faults) and to make internal node assignments. We present two extensions to EST. The first ensures that each portion of the ATPG search space is explored only once. The second applies headline objectives in parallel, rather than serially. For the 1965 ISCAS combinational benchmarks, EST accelerates S 0 CRATES by 6.53 times, when all faults are targeted, and by 5.51 times, when used with random pattern generation, f,iult simulation and fault dropping. This acceleration was achieved with minimal memory overhead.