Mengwei Xu;Daliang Xu;Chiheng Lou;Li Zhang;Gang Huang;Xin Jin;Xuanzhe Liu
{"title":"在 SoC 集群边缘服务器上进行高效、可扩展和可持续的 DNN 训练","authors":"Mengwei Xu;Daliang Xu;Chiheng Lou;Li Zhang;Gang Huang;Xin Jin;Xuanzhe Liu","doi":"10.1109/TMC.2024.3442430","DOIUrl":null,"url":null,"abstract":"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 \n<monospace>SoCFlow+</monospace>\n, a pioneering framework designed to facilitate the co-location of deep learning training tasks on SoC-Cluster servers, thereby optimizing resource utilization. \n<monospace>SoCFlow+</monospace>\n 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, \n<monospace>SoCFlow+</monospace>\n 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, \n<monospace>SoCFlow+</monospace>\n 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, \n<monospace>SoCFlow+</monospace>\n 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.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"23 12","pages":"14344-14360"},"PeriodicalIF":7.7000,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers\",\"authors\":\"Mengwei Xu;Daliang Xu;Chiheng Lou;Li Zhang;Gang Huang;Xin Jin;Xuanzhe Liu\",\"doi\":\"10.1109/TMC.2024.3442430\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"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 \\n<monospace>SoCFlow+</monospace>\\n, a pioneering framework designed to facilitate the co-location of deep learning training tasks on SoC-Cluster servers, thereby optimizing resource utilization. \\n<monospace>SoCFlow+</monospace>\\n 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, \\n<monospace>SoCFlow+</monospace>\\n 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, \\n<monospace>SoCFlow+</monospace>\\n 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, \\n<monospace>SoCFlow+</monospace>\\n 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.\",\"PeriodicalId\":50389,\"journal\":{\"name\":\"IEEE Transactions on Mobile Computing\",\"volume\":\"23 12\",\"pages\":\"14344-14360\"},\"PeriodicalIF\":7.7000,\"publicationDate\":\"2024-08-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE Transactions on Mobile Computing\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://ieeexplore.ieee.org/document/10634823/\",\"RegionNum\":2,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, INFORMATION SYSTEMS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Mobile Computing","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10634823/","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}
Efficient, Scalable, and Sustainable DNN Training on SoC-Clustered Edge Servers
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.
期刊介绍:
IEEE Transactions on Mobile Computing addresses key technical issues related to various aspects of mobile computing. This includes (a) architectures, (b) support services, (c) algorithm/protocol design and analysis, (d) mobile environments, (e) mobile communication systems, (f) applications, and (g) emerging technologies. Topics of interest span a wide range, covering aspects like mobile networks and hosts, mobility management, multimedia, operating system support, power management, online and mobile environments, security, scalability, reliability, and emerging technologies such as wearable computers, body area networks, and wireless sensor networks. The journal serves as a comprehensive platform for advancements in mobile computing research.