Haichen Shen, Lequn Chen, Yuchen Jin, Liangyu Zhao, Bingyu Kong, Matthai Philipose, A. Krishnamurthy, Ravi Sundaram
{"title":"Nexus: a GPU cluster engine for accelerating DNN-based video analysis","authors":"Haichen Shen, Lequn Chen, Yuchen Jin, Liangyu Zhao, Bingyu Kong, Matthai Philipose, A. Krishnamurthy, Ravi Sundaram","doi":"10.1145/3341301.3359658","DOIUrl":null,"url":null,"abstract":"We address the problem of serving Deep Neural Networks (DNNs) efficiently from a cluster of GPUs. In order to realize the promise of very low-cost processing made by accelerators such as GPUs, it is essential to run them at sustained high utilization. Doing so requires cluster-scale resource management that performs detailed scheduling of GPUs, reasoning about groups of DNN invocations that need to be co-scheduled, and moving from the conventional whole-DNN execution model to executing fragments of DNNs. Nexus is a fully implemented system that includes these innovations. In large-scale case studies on 16 GPUs, when required to stay within latency constraints at least 99% of the time, Nexus can process requests at rates 1.8-12.7X higher than state of the art systems can. A long-running multi-application deployment stays within 84% of optimal utilization and, on a 100-GPU cluster, violates latency SLOs on 0.27% of requests.","PeriodicalId":331561,"journal":{"name":"Proceedings of the 27th ACM Symposium on Operating Systems Principles","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2019-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"151","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 27th ACM Symposium on Operating Systems Principles","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3341301.3359658","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 151
Abstract
We address the problem of serving Deep Neural Networks (DNNs) efficiently from a cluster of GPUs. In order to realize the promise of very low-cost processing made by accelerators such as GPUs, it is essential to run them at sustained high utilization. Doing so requires cluster-scale resource management that performs detailed scheduling of GPUs, reasoning about groups of DNN invocations that need to be co-scheduled, and moving from the conventional whole-DNN execution model to executing fragments of DNNs. Nexus is a fully implemented system that includes these innovations. In large-scale case studies on 16 GPUs, when required to stay within latency constraints at least 99% of the time, Nexus can process requests at rates 1.8-12.7X higher than state of the art systems can. A long-running multi-application deployment stays within 84% of optimal utilization and, on a 100-GPU cluster, violates latency SLOs on 0.27% of requests.