LLM-Enhanced SoC Power Profiling and Predictive Energy Optimization for Heterogeneous Computing Systems
Main Article Content
Abstract
The growing adoption of heterogeneous computing systems, including multi-core System-on-Chip (SoC) architectures, graphics processors, cloud services, and diverse software ecosystems, has created significant challenges in power management and energy efficiency. This paper presents an LLM-enhanced framework for SoC power profiling and predictive energy optimization that combines operating system telemetry, hardware performance counters, application-level metrics, and cloud-based analytics to deliver comprehensive power insights. The proposed approach leverages Large Language Models (LLMs) to analyze and interpret power consumption patterns generated by workloads developed in Java, Python, C++, and JavaScript across Linux-based environments such as Fire OS and Vega OS. By integrating machine learning-based prediction techniques with LLM-driven reasoning, the framework identifies energy-intensive processes, forecasts future power demands, and recommends adaptive optimization strategies for CPU, GPU, memory, storage, and graphics subsystems. Experimental analysis demonstrates improved energy efficiency, enhanced resource utilization, and reduced power consumption while maintaining system performance. The findings highlight the potential of combining advanced AI techniques with system-level power management to enable intelligent, scalable, and sustainable computing platforms for next-generation embedded, edge, and cloud-connected applications.
Downloads
Article Details
Section
References
Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G. S., Davis, A., Dean, J., Devin, M., Ghemawat, S., Goodfellow, I., Harp, A., Irving, G., Isard, M., Jia, Y., Jozefowicz, R., Kaiser, Ł., Kudlur, M., ... Zheng, X. (2016). TensorFlow: Large-scale machine learning on heterogeneous distributed systems. arXiv. https://arxiv.org/abs/1603.04467
Henderson, P., Hu, J., Romoff, J., Brunskill, E., Jurafsky, D., & Pineau, J. (2020). Towards the systematic reporting of the energy and carbon footprints of machine learning. Journal of Machine Learning Research, 21(248), 1–43.
Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L., Rothchild, D., So, D., Texier, M., & Dean, J. (2021). Carbon emissions and large neural network training. arXiv. https://arxiv.org/abs/2104.10350
Wu, C., Raghavendra, R., Gupta, U., Acun, B., Ardalani, N., Maeng, K., Chang, G., Aga, M., Huang, J., Bai, C., Tian, M., & Wu, H. (2022). Sustainable AI: Environmental implications, challenges and opportunities. Proceedings of Machine Learning and Systems, 4, 795–813.
Luccioni, A. S., Viguier, S., & Ligozat, A. L. (2023). Estimating the carbon footprint of BLOOM, a large language model. Journal of Machine Learning Research, 24(253), 1–15.
Samsi, S., Zhao, J., McDonald, S., Li, B., Michaleas, P., Jones, M., Bergeron, W., Kepner, J., Tiwari, D., & Gadepally, V. (2023). From words to watts: Benchmarking the energy costs of large language model inference. arXiv. https://arxiv.org/abs/2310.03003
Patel, P., Choukse, E., Zhang, C., Goiri, Í., Lottarini, A., & Bianchini, R. (2024). Characterizing power management opportunities for large language models in the cloud. Proceedings of ASPLOS 2024, 1006–1023.
Wilkins, G., Keshav, S., & Mortier, R. (2024). Offline energy-optimal LLM serving: Workload-based energy models for LLM inference on heterogeneous systems. arXiv. https://arxiv.org/abs/2407.04014
Desislavov, R., Martínez-Plumed, F., & Hernández-Orallo, J. (2023). Compute and energy consumption trends in machine learning inference. Journal of Machine Learning Research, 24(125), 1–44.
Stone, J. E., McGaffey, A., Phillips, J. C., & Schulten, K. (2016). Evaluation of emerging energy-efficient heterogeneous computing platforms for scientific applications. Proceedings of the IEEE International Parallel and Distributed Processing Symposium Workshops, 951–960.
Zeng, Q., Du, Y., Huang, K., & Leung, K. K. (2020). Energy-efficient resource management for federated edge learning with CPU-GPU heterogeneous computing. IEEE Transactions on Wireless Communications, 20(4), 2447–2462.
Radovanović, A., Koningstein, R., Schneider, I., Chen, B., Duarte, A., Roy, B., Xiao, D., Haridasan, N., Hung, P., Care, N., & others. (2022). Carbon-aware computing for datacenters. IEEE Transactions on Power Systems, 38(2), 1270–1280.
Anderson, L., Berriel, R., Chockler, H., & Gomes, J. (2023). Energy-efficient deployment strategies for large language models. IEEE Access, 11, 118442–118456.
Dean, J., & Ghemawat, S. (2008). MapReduce: Simplified data processing on large clusters. Communications of the ACM, 51(1), 107–113.
Khan, S. U., Ahmad, I., & Kim, K. (2015). Energy-aware resource allocation in heterogeneous computing systems. Future Generation Computer Systems, 50, 77–89.
Mittal, S. (2014). A survey of techniques for improving energy efficiency in embedded computing systems. International Journal of Computer Aided Engineering and Technology, 6(4), 440–459.
Borkar, S., & Chien, A. A. (2011). The future of microprocessors. Communications of the ACM, 54(5), 67–77.
Hennessy, J. L., & Patterson, D. A. (2019). Computer architecture: A quantitative approach (6th ed.). Morgan Kaufmann.
Barroso, L. A., Clidaras, J., & Hölzle, U. (2018). The datacenter as a computer: Designing warehouse-scale machines (3rd ed.). Morgan & Claypool.
Grama, A., Gupta, A., Karypis, G., & Kumar, V. (2020). Introduction to parallel computing (3rd ed.). Pearson.
Pathak, S., Balantrapu, S. S., & Janakiraman, A. (2025). Future-Proofing the Planet: AI and XR for a Sustainable Tomorrow. In Exploring the Impact of Extended Reality (XR) Technologies on Promoting Environmental Sustainability (pp. 313-332). Cham: Springer Nature Switzerland.
Janakiraman, A. (2025). Explainability and Interpretability in Generative AI Agents. International Journal of Science, Technology and Convergence, 7(7).
Janakiraman, A. (2025). Muti-agent Generative Systems in E-commerce recommendations and pricing. Australian Journal of Cross-Disciplinary Innovation, 7(7).
Janakiraman, A. (2025). Leveraging Machine Learning for Equitable Green Innovation. In Advancing Social Equity Through Accessible Green Innovation (pp. 351-372). IGI Global Scientific Publishing.