Sarcouncil Journal of Multidisciplinary

Abstract: Accurate power modeling and estimation for smartphone System-on-Chips demands sophisticated correlation between pre-silicon predictions and post-silicon measurements. Modern smartphones integrate diverse functionalities within stringent power constraints, necessitating increasingly precise power prediction methodologies throughout the design cycle. While pre-silicon estimation provides critical early guidance, achieving high fidelity requires systematic calibration with actual silicon behavior. The technical review explores comprehensive strategies spanning architectural power modeling frameworks, component-level characterization techniques, and innovative correlation methodologies. Particular attention is given to the challenges inherent in modeling heterogeneous SoC architectures incorporating specialized accelerators alongside traditional processing elements. The integration of machine learning approaches with physics-based models demonstrates promising results for capturing complex power behaviors across diverse operating conditions. Through systematic post-silicon validation and iterative refinement processes, designers can develop highly accurate power models that enable optimal thermal management, extended battery life, and enhanced user experience in modern smartphone platforms. These techniques form the foundation for next-generation mobile devices that balance computational capability with energy efficiency, addressing the growing demands for performance within increasingly constrained power envelopes.