An Online Learning Methodology for Performance Modeling of Graphics Processors

• Published in: IEEE transactions on computers Vol. 67; no. 12; pp. 1677 - 1691
• Main Authors: Gupta, Ujjwal, Babu, Manoj, Ayoub, Raid, Kishinevsky, Michael, Paterna, Francesco, Gumussoy, Suat, Ogras, Umit Y.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.12.2018; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation models; Computational modeling; Distance learning; Error detection; Frames per second; frequency scaling; Graphics boards; Graphics processing units; Heuristic algorithms; Integrated GPUs; Machine learning; Mathematical models; Microprocessors; online learning; performance modeling; Power consumption; Power management; Power system management; Predictive models; RLS; Run time (computers); Sensitivity analysis; Smartphones; Time-frequency analysis; Weight reduction; Workloads
• ISSN: 0018-9340; 1557-9956
• DOI: 10.1109/TC.2018.2840710

Approximately 18 percent of the 3.2 million smartphone applications rely on integrated graphics processing units (GPUs) to achieve competitive performance. Graphics performance, typically measured in frames per second, is a strong function of the GPU frequency, which in turn has a significant impact on mobile processor power consumption. Consequently, dynamic power management algorithms have to assess the performance sensitivity to the frequency accurately to choose the operating frequency of the GPU effectively. Since the impact of GPU frequency on performance varies rapidly over time, there is a need for online performance models that can adapt to varying workloads. This paper presents a light-weight adaptive runtime performance model that predicts the frame processing time of graphics workloads at runtime without apriori characterization. We employ this model to estimate the frame time sensitivity to the GPU frequency, i.e., the partial derivative of the frame time with respect to the GPU frequency. The proposed model does not rely on any parameter learned offline. Our experiments on commercial platforms with common GPU benchmarks show that the mean absolute percentage error in frame time and frame time sensitivity prediction are 4.2 and 6.7 percent, respectively.