arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.01313cs.DCcs.LG
收藏

大规模GPU数据中心中深度学习工作负载的表征和预测

Characterization and Prediction of Deep Learning Workloads in Large-Scale GPU Datacenters

Nanyang Technological University(南洋理工大学);SenseTime(商汤科技)

Qinghao Hu, Peng Sun, Shengen Yan, Yonggang Wen, Tianwei Zhang

摘要:

现代GPU开发中心对于在研究社区和行业中提供深度学习(DL)模型和服务至关重要。在运营数据中心时,资源调度和管理的优化可以带来显著的经济效益。实现这一目标需要深入了解作业功能和用户行为。我们提出了一个全面的研究DL作业和资源管理的特点。首先,我们对SenseTime的真实工作轨迹进行了大规模分析。从集群、作业和用户的角度,我们发现了一些有趣的结论,这些结论可以促进集群系统的设计。其次,我们引入了一个通用框架,它基于历史数据管理资源。作为案例研究,我们设计:一个准最短的服务优先调度服务,它可以最大限度地减少集群范围内的平均作业完成时间高达6.5倍;和集群节能服务,提高整体集群利用率高达13%。

英文摘要:

Modern GPU datacenters are critical for delivering Deep Learning (DL) models and services in both the research community and industry. When operating a datacenter, optimization of resource scheduling and management can bring significant financial benefits. Achieving this goal requires a deep understanding of the job features and user behaviors. We present a comprehensive study about the characteristics of DL jobs and resource management. First, we perform a large-scale analysis of real-world job traces from SenseTime. We uncover some interesting conclusions from the perspectives of clusters, jobs and users, which can facilitate the cluster system designs. Second, we introduce a general-purpose framework, which manages resources based on historical data. As case studies, we design: a Quasi-Shortest-Service-First scheduling service, which can minimize the cluster-wide average job completion time by up to 6.5x; and a Cluster Energy Saving service, which improves overall cluster utilization by up to 13%.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑