NIC-based offload of dynamic user-defined modules for Myrinet clusters

2004 IEEE International Conference on Cluster Computing (IEEE Cat. No.04EX935) Pub Date : 2004-09-20 DOI:10.1109/CLUSTR.2004.1392618

A. Wagner, Hyun-Wook Jin, D. Panda, R. Riesen

{"title":"NIC-based offload of dynamic user-defined modules for Myrinet clusters","authors":"A. Wagner, Hyun-Wook Jin, D. Panda, R. Riesen","doi":"10.1109/CLUSTR.2004.1392618","DOIUrl":null,"url":null,"abstract":"Many of the modern networks used to interconnect nodes in cluster-based computing systems provide network-interface cards (NICs) that offer programmable processors. Substantial research has been done with the focus of offloading processing from the host to the NIC processor. However, the research has primarily focused on the static offload of specific features to the NIC, mainly to support the optimization of common collective and synchronization-based communications. We describe the design and implementation of a framework based on MP1CH-GM to support the dynamic NIC-based offload of user-defined modules for Myrinet clusters. We evaluate our implementation on a 16-node cluster using a NIC-based version of the common broadcast operation and we find a maximum factor of improvement of 1.2 with respect to total latency as well as a maximum factor of improvement of 2.2 with respect to average CPU utilization under conditions of process skew. In addition, we see that these improvements increase with system size, indicating that our NIC-based framework offers enhanced scalability when compared to a purely host-based approach.","PeriodicalId":123512,"journal":{"name":"2004 IEEE International Conference on Cluster Computing (IEEE Cat. No.04EX935)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2004-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"22","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2004 IEEE International Conference on Cluster Computing (IEEE Cat. No.04EX935)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CLUSTR.2004.1392618","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 22

Abstract

Many of the modern networks used to interconnect nodes in cluster-based computing systems provide network-interface cards (NICs) that offer programmable processors. Substantial research has been done with the focus of offloading processing from the host to the NIC processor. However, the research has primarily focused on the static offload of specific features to the NIC, mainly to support the optimization of common collective and synchronization-based communications. We describe the design and implementation of a framework based on MP1CH-GM to support the dynamic NIC-based offload of user-defined modules for Myrinet clusters. We evaluate our implementation on a 16-node cluster using a NIC-based version of the common broadcast operation and we find a maximum factor of improvement of 1.2 with respect to total latency as well as a maximum factor of improvement of 2.2 with respect to average CPU utilization under conditions of process skew. In addition, we see that these improvements increase with system size, indicating that our NIC-based framework offers enhanced scalability when compared to a purely host-based approach.

查看原文本刊更多论文

为Myrinet集群提供基于网卡的动态用户定义模块卸载

许多用于在基于集群的计算系统中互连节点的现代网络都提供了提供可编程处理器的网络接口卡(nic)。大量的研究集中在将处理从主机卸载到NIC处理器上。然而，研究主要集中在特定功能的静态卸载到NIC上，主要是为了支持公共集体和基于同步的通信的优化。我们描述了一个基于MP1CH-GM的框架的设计和实现，以支持Myrinet集群中用户定义模块的基于nic的动态卸载。我们使用基于nic的公共广播操作版本来评估我们在16节点集群上的实现，我们发现在进程倾斜的条件下，关于总延迟的最大改进因子为1.2，关于平均CPU利用率的最大改进因子为2.2。此外，我们看到这些改进随着系统大小的增加而增加，这表明与纯粹基于主机的方法相比，我们基于nic的框架提供了增强的可伸缩性。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

2004 IEEE International Conference on Cluster Computing (IEEE Cat. No.04EX935)

自引率

0.00%

发文量