CISE Research Infrastructure: Network of Symmetric Multiprocessors: An Infrastructure for Research in Scalable Systems and Applications

CISE 研究基础设施：对称多处理器网络：可扩展系统和应用研究的基础设施

基本信息

批准号：
9624099
负责人：
Kai Li
金额：
$ 153.37万
依托单位：
Princeton University
依托单位国家：
美国
项目类别：
Continuing Grant
财政年份：
1996
资助国家：
美国
起止时间：
1996-09-01 至 2001-08-31
项目状态：
已结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=9624099&HistoricalAwards=false
关键词：
CISE Research Infrastructure Network Symmetric

项目摘要

9624099 Li, Kai Dobkin, David Princeton University CISE Research Infrastructure: Network of Symmetric Multiprocessors : An Infrastructure for Research in Scalable Systems and Applications This award provides support for a research infrastructure based on commodity PC-based Symmetric MultiProcessor (SMP) servers, high-speed networks, and customized network interfaces and communication software. It will enable new research activities, extension of on-going research activities and will generate synergy among the several research efforts. Key research areas which will be enabled by the infrastructure include virtual memory-mapped communication, efficient shared virtual memory, scalable virtual storage servers, efficient message passing, and performance measurement. Application areas include scalable DNA computing simulations, parallel fluid dynamics, particle-based computation in a homogeneous medium and projects in parallel computational biology. The infrastructure consists primarily of stock hardware, but in quantity large enough to investigate the viability of building a large-scale scalable system. The infrastructure will be constructed in three phases: an initial phase, a building phase, and a large-scale experimental phase. In the initial phase, four PC-based SMP servers will be connected together with a custom-designed network. In the building phase, 16 PC based SMP servers will be connected with a high speed commercial network and a Myricom/Princeton customized network. In the experimental phase, 32 next generation PC-based SMP servers will be connected with a commercially available high-speed network and a Myricom/Princeton customized network. All research projects will utilize the infrastructure to conduct large-scale experiments.

9624099 LI，Kai Dobkin，David Princeton University CISE研究基础设施：对称多处理器网络：可扩展系统和应用研究的基础架构该奖项为基于商品PC的基于PC的对称对称性对称性处理器（SMP）服务器（SMP）服务器，高速网络和定制的软件和定制的软件的研究基础架构提供支持。它将实现新的研究活动，扩展正在进行的研究活动，并在几项研究工作中产生协同作用。基础架构将启用的关键研究领域包括虚拟内存映射通信，有效的共享虚拟内存，可扩展的虚拟存储服务器，有效的消息传递和性能测量。应用领域包括可扩展的DNA计算模拟，平行流体动力学，均匀介质中的基于粒子的计算以及平行计算生物学中的项目。基础设施主要由库存硬件组成，但数量足够大，可以研究建立大规模可扩展系统的可行性。基础设施将分为三个阶段：初始阶段，建筑阶段和大规模的实验阶段。在初始阶段，将将四个基于PC的SMP服务器与自定义设计的网络一起连接。在建筑阶段，将与高速商业网络和MyRicom/Princeton自定义网络连接16个基于PC的SMP服务器。在实验阶段，将与32个基于PC的SMP服务器与市售高速网络和MyRicom/Princeton自定义网络连接。所有研究项目都将利用基础设施进行大规模实验。