干旱区生态环境知识资源中心(Arid): Topology-Aware Job Scheduling for Machine Learning Cluster

Arid

	Topology-Aware Job Scheduling for Machine Learning Cluster
	Lu, Jingyuan; Li, Peng; Wang, Kun; Feng, Huibin; Guo, Enting; Wang, Xiaoyan; Guo, Song
通讯作者	Lu, JY (corresponding author), Nanjing Univ Posts & Telecommun, Nanjing, Peoples R China.
会议名称	IEEE Global Communications Conference (GLOBECOM)
会议日期	DEC 09-13, 2019
会议地点	Waikoloa, HI
英文摘要	Parameter Server (PS) has been widely used to train a large amount of data on multiple machines in parallel. In parameter server, a critical problem is how to effectively schedule multiple training jobs to minimize the job completion time. Some existing work has proposed methods of setting the number of concurrent workers. However, they do not effectively consider the topology of GPU placement which affects the efficiency of communication. This paper proposes a novel resource-to-time model based on the number of workers and the topology of GPU placement. According to the model, we propose an algorithm called TOPO-PS particularly for topology problem in parameter servers. The algorithm achieves the placement strategy based on graph mapping algorithm. Evaluation under various algorithms evidences the superiority of our algorithm. TOPO-PS yields shorter job completion, by up to 53.48% of that of FIFO and 88.77% of OASIS.
英文关键词	Parameter Server Scheduling Algorithms Cloud Computing Machine Learning
来源出版物	2019 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM)
ISSN	2334-0983
出版年	2019
ISBN	978-1-7281-0962-6
出版者	IEEE
类型	Proceedings Paper
语种	英语
收录类别	CPCI-S
WOS记录号	WOS:000552238601101
WOS关键词	ALLOCATION
WOS类目	Computer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS研究方向	Computer Science ; Engineering ; Telecommunications
资源类型	会议论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/370059
作者单位	[Lu, Jingyuan; Guo, Enting] Nanjing Univ Posts & Telecommun, Nanjing, Peoples R China; [Li, Peng] Univ Aizu, Aizu Wakamatsu, Fukushima, Japan; [Feng, Huibin] Minjiang Univ, Fuzhou, Peoples R China; [Wang, Kun] Univ Calif Los Angeles, Los Angeles, CA USA; [Wang, Xiaoyan] Ibaraki Univ, Ibaraki, Japan; [Guo, Song] Hong Kong Polytech Univ, Hong Kong, Peoples R China
推荐引用方式 GB/T 7714	Lu, Jingyuan,Li, Peng,Wang, Kun,et al. Topology-Aware Job Scheduling for Machine Learning Cluster[C]:IEEE,2019.