Performance-Aware Energy-Efficient GPU Frequency Selection using DNN-based Models


Energy efficiency will be important in future accelerator-based HPC systems for sustainability and to improve overall performance. This study proposes a deep neural network (DNN)-based learning model for execution time and power consumption of workloads across GPUs DVFS design space. Micro-architectural data obtained by running SPEC-ACCEL, DGEMM, and STREAM benchmarks are used for model training. These features are consistent for a workload unaffected by frequency and input size reducing the data required significantly. For real-world applications - LAMMPS, NAMD, GROMACS, LSTM, BERT, and ResNet50 power and time models show 89% - 98% accuracy on NVIDIA Ampere. Multi-objective functions help select optimal frequencies that lower power and minimize performance impact showing maximum energy savings of 27% at a performance loss of 1.8%. The same models trained on Ampere showed an accuracy of greater than 93% on an NVIDIA Volta, thereby demonstrating model portability across architectures.


© 2023 Association for Computing Machinery. All rights reserved. cc-by


Ampere GPU, dynamic voltage frequency scaling, energy-efficiency, GPU, high-performance computing, Volta GPU


Ali, G., Side, M., Bhalachandra, S., Wright, N.J., & Chen, Y.. 2023. Performance-Aware Energy-Efficient GPU Frequency Selection using DNN-based Models. ACM International Conference Proceeding Series.