University of Texas at Arlington
Existing power management techniques operate by reducing performance capacity (frequency, voltage, resource size) when performance demand is low, such as at idle or similar low activity phases. In the case of multi-core systems, the performance and power demand is the aggregate demand of all cores in the system. Monitoring aggregate demand makes detection of phase changes difficult (active-to-idle, idle-to-active, etc.) since aggregate phase behavior obscures the underlying phases generated by the workloads on individual cores. This causes sub-optimal power management and over-provisioning of power resources.