中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
Sketch-fusion: A gradient compression method with multi-layer fusion for communication-efficient distributed training

文献类型:期刊论文

作者Dai, Lingfei1,2; Gong, Luqi1; An, Zhulin1; Xu, Yongjun1; Diao, Boyu1
刊名JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING
出版日期2024-03-01
卷号185页码:10
关键词Gradient compression Multi-layer fusion Distributed stochastic gradient descent Deep learning training
ISSN号0743-7315
DOI10.1016/j.jpdc.2023.104811
英文摘要Gradient compression is an effective technique for improving the efficiency of distributed training. However, introducing gradient compression can reduce model accuracy and training efficiency. Furthermore, we also find that using a layer-wise gradient compression algorithm would lead to significant compression and communication overhead, which can negatively impact the scaling efficiency of the distributed training system. To address these issues, we propose a new method called Sketch-Fusion SGD, which leverages the Count-Sketch data structure to enhance the scalability and training speed of distributed deep learning systems. Moreover, our method employs LayerFusion to optimize gradient compression algorithms' scalability and convergence efficiency by formulating an optimal multi-layer fusion strategy without introducing extra hyperparameters. We evaluate our method on a cluster of 16 GPUs and demonstrate that it can improve training efficiency by up to 18.6% without compromising the model's accuracy. In addition, we find that applying our LayerFusion algorithm to other gradient compression methods improved their scalability by up to 2.87x.
WOS研究方向Computer Science
语种英语
WOS记录号WOS:001127654600001
出版者ACADEMIC PRESS INC ELSEVIER SCIENCE
源URL[http://119.78.100.204/handle/2XEOYT63/38453]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Diao, Boyu
作者单位1.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China
2.Univ Chinese Acad Sci, Coll Comp Sci, Beijing, Peoples R China
推荐引用方式
GB/T 7714
Dai, Lingfei,Gong, Luqi,An, Zhulin,et al. Sketch-fusion: A gradient compression method with multi-layer fusion for communication-efficient distributed training[J]. JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING,2024,185:10.
APA Dai, Lingfei,Gong, Luqi,An, Zhulin,Xu, Yongjun,&Diao, Boyu.(2024).Sketch-fusion: A gradient compression method with multi-layer fusion for communication-efficient distributed training.JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING,185,10.
MLA Dai, Lingfei,et al."Sketch-fusion: A gradient compression method with multi-layer fusion for communication-efficient distributed training".JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING 185(2024):10.

入库方式: OAI收割

来源:计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。