Grant: Collaborative Research and Development Grants

This project builds upon two areas of ongoing collaboration between Intel and the University of British Columbia. The first topic, formal verification uses mathematical techniques to establish critical properties of designs and avoid costly errors. The second is the development of novel circuit techniques to address challenges in deep-submicron design. This project brings these two areas together in the design of active-inductor circuits and the demonstration of formal methods for analog circuit design. Inductors play an important role in the interconnect between chips in a larger circuit. The traditional implementation of an inductor using a coil of wire has serious drawbacks when used on-chip. Such passive inductors consume large areas and have severe losses. Active inductors replace the coil of wire with circuits composed of transistors and capacitors that mimic an inductor's behaviour. Such circuits take good advantage of the small transistors afforded by advanced fabrication processes and provide a further benefit of programmability. However, many issues including robustness, power consumption, and high-frequency limitations still remain. This project addresses these issues. Formal methods based on discrete models have achieved significant success. This project extends the applicability of formal methods to analog circuits and high-performance digital designs. This project will identify benchmark verification problems, develop a common front-end to be used with several recently developed verification tools, and evaluate these tools using the benchmark circuits. Circuits arising from the active inductor research will be considered along with other examples from high-performance digital design. This process will guide further development of the tools and identify opportunities for incorporating these approaches into commercial design methods.