Dynamic power management of systems

Abstract

Abstract A method and system for power management of components within a system by introducing low-level instructions, such as NOPS, delay loops, and sleep modes, into the instruction sequences within logical components of the system, thereby allowing the system to maintain desirable power dissipation over a given time interval. Through employment of a control mechanism, a system can interrupt itself periodically to determine if the power dissipation must be limited and to what extent. A system can determine if and by how much it must limit its power by polling external components. The method may be implemented by a controller that intelligently manages power dissipation of components through selective manipulation of computer operations by selected components. Controll of power dissipation for the selected components is intelligently based on overall system performance requirements. Furthermore, a plurality of sensors can measure power dissipation of a plurality of components in the system and cause the controller to selectively manipulate those components that require power management. In a proactive approach, power dissipation for the plurality of managed components can be controlled based on known power characteristics for the selected components without necessarily monitoring the power dissipation of the system through sensors. In a reactive approach, a plurality of sensors that measure power dissipation in the system can cause the controller to selectively manipulate components requiring power management. The system can employ a combination of proactive and reactive approaches to power management.

Inventors

• 👤 Mark H. Linderman

Bibliographic Details