A Novel Path Selection Method for Small Delay Defects Test

662 Words2 Pages

Aggressive technology scaling has resulted in increasing process variations and statistical diversity in manufacturing. Process variations result in varying in path lengths, and thus a possibly different set of critical paths for different process corners, necessitating consideration of process variation in delay test methods. Also, process variation adds some of near-critical paths to the critical/longest paths set [4]. Therefore, to maintain the reliability of circuits, testing methodologies need to be improved. As technologies scales down process variation effect become more significant. Process variation causes circuit performance deviate from its initial design expectation. Therefore, it may lead to reducing the timing margins and hence, increasing the timing violation probability. As the result by shrinking the size of the transistors the importance of delay test has become more and more [10]. Process variation is a combination of systematic effects and random effects (e.g. the number of dopant atoms implanted in a transistor) which cause variation in frequency [9]. It should be noted that the random type variations especially RDF (Random Dopant Fluctuation) are dominant [14]. In order to deal with timing defects, two fault models had been used, path delay fault and transition delay fault. Small delay defects (SDD) are one type of delay defects which is affected by variations like process variation, cross-talk and power supply noise effects. SDDs introduce a small amount of extra delay to design [10] in new technology scales. The delay introduced by SDD is small, but the overall impact can effect on the performance of the target circuit if the sensitized path is critical. In order to capture cumulative effect of small delay... ... middle of paper ... ...proposed method, by considering the gates sharing between the paths, prunes the candidate critical path set U. All the paths in the U have the potential to violate the predefined timing constraint (TC). Pruning the path set before provides us to use optimal methods with an acceptable runtime. Therefore, after this phase, we propose to use an Integer Linear Programming (ILP) method to find the best path set from the pruned set. For ILP selection, we propose an objective function which considers both correlation between paths and, the criticality of each path. The rest of this paper is organized as follows. The works in the field of the path selection on the presentence of the process variation are addressed in Section II. Section III describes the proposed heuristic method. The experimental results are described in section IV and the paper is concluded in Section V.

Open Document