摘要：Computational sprinting was introduced to tackle the "dark silicon" problem, i.e., to allow a processor to transiently consume a power that could not be sustained indefinitely without thermal damage. However, the idea of sprinting has other potential applications, also tightly related to embedded systems. In this paper we evidence a few of these, and discuss thescenarioin a view to abstracting and defining general problems. We then propose an event-based approach to treat the said problems in a unitary manner.
