Flexible Architectural Cores for Commerical Computing Systems: A University-Industry GOALI Collaboration

用于商业计算系统的灵活架构核心：大学与行业 GOALI 合作

• 批准号: 9900605
• 负责人: David Lilja
• 金额: $ 19.69万
• 依托单位: University of Minnesota-Twin Cities
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1999
• 资助国家: 美国
• 起止时间: 1999-09-15 至 2003-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9900605&HistoricalAwards=false
• 关键词: Flexible; Architectural; Cores; Commerical; Computing

Commercial computing workloads, such as on-line transactionprocessing, file-serving, data mining, and web-serving, currentlyrepresent about 85 percent of the overall market for large computingsystems. Nevertheless, much of the computer architecture researchto date has focused on improving the capability of computer systemsto access, process, and store the large arrays of data common tonumerically-intensive workloads. The implicit assumption has beenthat all types of computer systems would benefit from thearchitectural innovations developed for these applications. Toaddress this disparity between the economic importance of commercialcomputing systems and the traditional architecture development focuson technical computing systems, this project is developing theFlexible Architectural Cores for Commercial Computing Systemsprocessor architecture. This new architecture is based on theobservation that the constantly changing behavior in a commercialcomputing environment causes different architectural innovations tobe applicable during different phases of execution. A unique aspectof this project that will significantly enhance its relevance tocommercial computing environments is the close collaboration of theIBM Corporation's commercial microprocessor development group inRochester, Minnesota. This project will make substantialcontributions to educational and human resource development throughgraduate student research and interaction with industry.

商业计算工作负载，如在线事务处理、文件服务、数据挖掘和Web服务，目前占大型计算系统整体市场的85%。 然而，迄今为止，大部分计算机体系结构研究都集中在提高计算机系统访问、处理和存储大量数据的能力上，这些数据通常用于数字密集型工作负载。 隐含的假设是，所有类型的计算机系统都将受益于为这些应用程序开发的架构创新。 为了解决商业计算系统的经济重要性与传统架构开发集中于技术计算系统之间的这种差异，本项目正在开发商业计算系统处理器架构的灵活架构核心。 这种新的体系结构是基于这样的观察，即商业计算环境中不断变化的行为导致不同的体系结构创新适用于不同的执行阶段。 这个项目的一个独特的方面，将显着提高其相关的商业计算环境是密切合作的IBM公司的商业微处理器开发组在罗切斯特，明尼苏达州。 该项目将通过研究生研究和与工业界的互动，为教育和人力资源开发做出实质性贡献。