Adapting asynchronous controllers to operating conditions by logic parametrization

Modern hardware systems are required to be robust, resilient and long-life, and thus they have to be adaptive to possibly changing requirements and operating conditions. This covers not only the data processing functions but also control, and even timing and power operation. For example, such systems will be increasingly powered by ambient sources (energy harvesting) and will experience a wide range of power modes, including not only dynamic voltage scaling but even power supplies which do not deliver a stable level of Vdd. Asynchronous controllers can o er ultimate robustness under extreme variations caused by unstable power supply; however, the robustness often comes at the price of a heavy and slow implementation. This implementation may be the only option for the extreme operating conditions, but is unacceptable for the normal operation mode of the system (in which it may be running most of the time). There is clearly a need to choose a particular controller implementation in run-time rather than in design time. In this paper we present and investigate a new methodology which provides a way of building adaptive asynchronous controllers supporting a range of power and timing modes of operation (possibly even synchronous modes) that can be selected during run-time according to the known hardware, environmental, and power/latency budget requirements.