Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers
IEEE Transactions on Mobile Computing · 2024-12 · 23 (12) · 14344-14360
DOI: 10.1109/tmc.2024.3442430 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Mobile Computing, 2024-12. Verfasst von Mengwei Xu und 6 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 94 Literaturangaben. Gefördert von: National Science and Technology Major Project of China, National Natural Science Foundation of China.
Abstract (Originalsprache)
In the realm of industrial edge computing, a novel server architecture known as SoC-Cluster, characterized by its aggregation of numerous mobile systems-on-chips (SoCs), has emerged as a promising solution owing to its enhanced energy efficiency and seamless integration with prevalent mobile applications. Despite its advantages, the utilization of SoC-Cluster servers remains unsatisfactory, primarily attributed to the tidal patterns of user-initiated workloads. To address such inefficiency, we introduce SoCFlow+, a pioneering framework designed to facilitate the co-location of deep learning training tasks on SoC-Cluster servers, thereby optimizing resource utilization. SoCFlow+ incorporates three novel techniques tailored to mitigate the inherent limitations of commercial SoC-Cluster servers. First, it employs group-wise parallelism complemented by delayed aggregation, a strategy engineered to enhance the training efficiency and scalability of deep learning models, effectively circumventing network bottlenecks. Second, it integrates a data-parallel mixed-precision training algorithm, optimized to exploit the heterogeneous processing capabilities inherent to mobile SoCs fully. Third, SoCFlow+ employs an underclocking-aware workload re-balanacing mechanism to tackle the training performance degradation caused by the thermal control of mobile SoCs. Through rigorous experimental validation, SoCFlow+ achieves a convergence speedup ranging from 1.6× to 740× across 32 SoCs, compared to conventional benchmarks. Furthermore, when juxtaposed with commodity GPU servers (e.g., NVIDIA V100) under identical power constraints, SoCFlow+ not only exhibits comparable training speed but also achieves a remarkable reduction in energy consumption by a factor of 2.31× to 10.23×, all while preserving convergence accuracy.
Förderung
- National Science and Technology Major Project of China
- National Natural Science Foundation of China
Zitieren
APA
Xu, M., Xu, D., Lou, C., Zhang, L., Huang, G., Jin, X., & Liu, X. (2024). Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers. IEEE Transactions on Mobile Computing, 23(12), 14344-14360. https://doi.org/10.1109/tmc.2024.3442430
Vancouver
Xu M, Xu D, Lou C, Zhang L, Huang G, Jin X, et al. Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers. IEEE Transactions on Mobile Computing. 2024;23(12):14344-14360. doi:10.1109/tmc.2024.3442430
BibTeX
@article{Xu2024,
author = {Xu, Mengwei and Xu, Daliang and Lou, Chiheng and Zhang, Li and Huang, Gang and Jin, Xin and Liu, Xuanzhe},
title = {Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers},
journal = {IEEE Transactions on Mobile Computing},
year = {2024},
volume = {23},
number = {12},
pages = {14344--14360},
doi = {10.1109/tmc.2024.3442430}
}
EUVIMED