Efficient and reliable fault analysis methodology for nanomagnetic circuits

The increasing issues in scaled Complementary Metal Oxide Semiconductor (CMOS) circuit fabrication favor the flourishing of emerging technologies. Because of their limited sizes, both CMOS and emerging technologies are particularly sensitive to defects that arise during the fabrication process. Their impact is not easy to analyze in order to take the necessary countermeasures, especially in the case of circuits of realistic complexity based on emerging technologies. In this work, we propose a new methodology supported by an efficient and reliable tool for the identification of the impact of faults in complex circuits implemented using the emerging technology we are focusing on in this case: nanomagnetic logic. The methodology is based on three main steps: (i) we performed exhaustive physical-level simulations of basic blocks based on a detailed finite-element tool in order to have a full characterization, to know their properties in presence of defects, and to have a solid reference point for the following steps; (ii) we developed a model ( FANOMAG ) for the basic block behavior suitable for simulations in presence of defects of complex circuits, that is, lighter than a physical level one, but accurate enough to capture the most impor- tant features to be inherited at circuit level; (iii) starting from a physical design of complex circuits that we perform using a specific design tool we developed, that is, TOPOLINANO , we simulated using FANOMAG , now embedded in our TOPOLINANO tool, the behavior of circuits in presence of multiple sets of fabrication defects using a Monte Carlo approach now included in TOPOLINANO as a new feature. In this paper, a specific type of defect is considered as a case study. The framework and methodology are conceived to be easily extended to handle other types of defects and problems due to working conditions that a designer and/or a technologist might want to focus on. The major outcome is then a powerful methodology and tool capable to analyze with a good accuracy nanomagnetic logic complex circuits and architectures both in ideal conditions and in presence of defects with remarkable performance in terms of simulation times.