On-chip thermal management techniques using inter-processor time dependent power density data for indentification of thermal aggressors

Abstract

Various embodiments of methods and systems for controlling and/or managing thermal energy generation on a portable computing device are disclosed. Data discarded from one or more processing core registers may be monitored and analyzed to deduce individual workloads that have been processed by each of the cores over a unit of time. From the deduced workloads, the power consumed by each of the cores over the unit of time in order to process the workload can be calculated. Subsequently, a time dependent power density map can be created which reflects a historical and near real time power consumption for each core. Advantageously, because power consumption can be correlated to thermal energy generation, the TDPD map can be leveraged to identify thermal aggressors for targeted, fine grained application of thermal mitigation techniques. In some embodiments, workloads may be reallocated from the identified thermal aggressors to the identified underutilized processing components.