Phase-Based Accurate Power Modeling for Mobile Application Processors

Modern mobile application processors are required to execute heavier workloads while the battery capacity is rarely increased. This trend leads need for a power model that can analyze consumed by CPU and GPU at run-time, which key components of processor in terms savings. We propose novel models based on phases using performance monitoring counters smartphones. Our phase-based employ combined per-phase modeling methods achieve more accurate consumption estimations, unlike existing models. The proposed shows estimation errors 2.51% ARM Cortex A-53 1.97% Samsung M1 average, an average error 8.92% Mali-T880. In addition, we integrate with latest display into holistic model. estimate smartphone′s total 6.36% running nine 3D game benchmarks, improving rate about 56% compared prior