Robust and Accurate Fine-Grain Power Models for Embedded Systems With No On-Chip PMU

This letter a novel approach to event-based power modeling for embedded platforms that do not have performance monitoring unit (PMU). The method involves complementing the target hardware platform, where physical data is measured, with another platform on which CPU data, needed model generation, can be collected. methodology used generate accurate fine-grain models Gaisler GR712RC dual-core LEON3 fault-tolerant SPARC processor onboard sensors and no PMU. A Kintex UltraScale field-programmable gate array (FPGA) as support obtain required by running soft-core representation of but PMU implementation. Both execute same benchmark set collection synchronized using per-sample timestamps so sensor from board matched FPGA. samples are then processed Robust Energy Power Predictor Selection (REPPS) software in order models. achieve less than 2% estimation error when validated an industrial use case follow program phases, makes them suitable runtime profiling during development.