Loading...

A Machine Learning Approach to Minimize Power Consumption of Smartphones While Satisfying the Gaming Performance

Aghapour, Ehsan | 2020

312 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53182 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Sarbazi Azad, Hamid
  7. Abstract:
  8. Today's smartphone devices include several cores, such as CPU, GPU, and different accelerators, in order to maximize user experience. However, due to meeting the power budget and limited capacity battery, power and energy of their cores should be managed using dynamic power management methods such as dynamic voltage and frequency scaling (DVFS). For this purpose, we should find optimal frequency and voltage settings of processing cores for each time, to minimize energy consumption while retaining user experience. Finding this optimal frequency and voltage settings is a challenging problem that depends on many parameters. We propose to use deep reinforcement learning (DRL) method to automatically learn near-optimal frequency settings of processing cores for each time, based on observed state of environment. We implement our proposed DRL on the neural network accelerator integrated into the Kirin970 smartphone SoC. We target game applications because of their high energy consumption and their popularity among smartphone users. For this purpose, we train the neural network with multiple game applications and then use the trained model as governor in the source code of the android and kernel. Using our method as governor for three games shows that our method can reduce total energy consumption of GPU and CPU cores, on average 30% while satisfying user experience, in comparison to the high-performance android governor
  9. Keywords:
  10. Smart Phones ; Power Consumption ; Machine Learning ; Dynamic Voltage and Ferquency Scaling (DVFS) ; Graphics Procssing Unit (GPU) ; User Experience

 Digital Object List

 Bookmark

No TOC