Energy consumption is a major concern in multicore systems. Perhaps thesimplest strategy for reducing energy costs is to use only as many cores asnecessary while still being able to deliver a desired quality of service.Motivated by earlier work on a dynamic (heterogeneous) core allocation schemefor H.264 video decoding that reduces energy costs while delivering desiredframe rates, we formulate operationally the general problem of executing asequence of actions on a reconfigurable machine while meeting a correspondingsequence of absolute deadlines, with the objective of reducing cost. Using atransition system framework that associates costs (e.g., time, energy) withexecuting an action on a particular resource configuration, we use the notionof amortised cost to formulate in terms of simulation relations appropriatenotions for comparing deadline-conformant executions. We believe these notionscan provide the basis for an operational theory of optimal cost executions andperformance guarantees for approximate solutions, in particular relating thenotion of simulation from transition systems to that of competitive analysisused for, e.g., online algorithms.
展开▼
