Application of memory scrambling aware multi-level diagnosis flow

To overcome the issues of safety, reliability and efficiency in nanoscale designs, advanced methods of fault detection and diagnosis were developed. Multi-level model based methods of fault detection were proposed using a hierarchy of detection and diagnosis methods and dynamic models, since previous approaches do not give a deeper insight and mainly limit or trend checking of some measurable output variables, which usually makes it impossible to do fault diagnosis. The new developed methods generate several symptoms indicating the difference between nominal and faulty statuses. Based on different symptoms, fault diagnosis procedures follow, determining the fault by applying the developed classification scheme. In this paper, the validity of a memory scrambling aware multi-level fault diagnosis flow is shown by experiments and different case scenarios.