A ResNet-Based DVFS Regulator for Heterogeneous Multi-Core Mobile Processors
Dynamic voltage and frequency scaling (DVFS) is commonly used for the balance of performance and power of mobile processors. Conventional DVFS regulators take the processor cores with the same power and performance weight showing their inability in the regulation of heterogeneous processors. This paper proposes a core-aware frequency-power evaluation scheme that utilizes a multilayer perceptron (MLP) to assess the efficacy of DVFS regulation actions. Additionally, a ResNet-based DVFS regulator, trained with the assistance of the MLP, is developed for heterogeneous multi-core processors. Evaluated on Xiaomi 13 Pro powered by a Qualcomm Snapdragon 8 Gen 2, the proposed DVFS approach achieves an up to ×1.47 improvement in FPS Performance Per Watt (FPPW) compared with the Qualcomm's default DVFS governor in video recording scenario.
Authors
- Yang Zhao (ORCID: https://orcid.org/0000-0002-4032-8049)
- Shibo Hu
- Yong Lian (ORCID: https://orcid.org/0000-0002-5289-5219)
- Yuze Chen (ORCID: https://orcid.org/0000-0001-8504-1458)
- Xinzi Xu (ORCID: https://orcid.org/0000-0001-5869-1631)
- Qin Mao (ORCID: https://orcid.org/0009-0008-6146-7794)
Institutions
- China Mobile (China) (CN)
- Shanghai Jiao Tong University (CN)
- York University (CA)
- Xiaomi (China) (CN)
Publication Details
- Published
- 2024-12-01
- DOI
- https://doi.org/10.1109/tencon61640.2024.10902794
- Primary Topic
- Advanced Neural Network Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00