Litcius/Paper detail

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

Ghazanfar Ali, Mert Side, Sridutt Bhalachandra, Nicholas J. Wright, Yong Chen

202318 citationsDOIOpen Access PDF

Abstract

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.

Topics & Concepts

Computer scienceSoftware portabilityWorkloadEfficient energy useDesign space explorationEnergy consumptionSpec#Benchmark (surveying)Parallel computingArtificial neural networkComputer engineeringPower (physics)Deep learningEnergy (signal processing)Computer architectureArtificial intelligenceEmbedded systemOperating systemProgramming languageElectrical engineeringEcologyBiologyPhysicsGeodesyEngineeringGeographyStatisticsMathematicsQuantum mechanicsParallel Computing and Optimization TechniquesAdvanced Data Storage TechnologiesAdvanced Neural Network Applications
Performance-Aware Energy-Efficient GPU Frequency Selection using DNN-based Models | Litcius