BRIGE: OS-SPECIFIC MANY-CORE DESIGN

BRIGE：特定于操作系统的多核设计

• 批准号: 1125762
• 负责人: Chen Liu
• 金额: $ 17.5万
• 依托单位: Florida International University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-09-01 至 2012-11-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1125762&HistoricalAwards=false
• 关键词: BRIGE; OS; SPECIFIC; MANY; CORE

Intellectual Merit: The objective of this BRIGE research project is to leverage current microprocessor design to support and meet the increasing demand on system performance/ throughput. The ever-increasing complexity of microprocessors makes it necessary to renovate the current technology. The project will explore the design space of many-core microprocessors, centered on reducing the operating system overhead and interference with user applications via hardware support. This project is expected to yield a more detailed understanding of the interactions between the hardware and system software that take part during the execution of applications. A transformative scheme is proposed to distribute the operating system services onto dedicated and specialized cores. The novelty of this research lies in an attempt to combine research fronts at different layers of the microarchitectural design into a unified framework, with specific focus on scalability, adaptivity, and improved fault handling. This project will advance the fundamental understanding of principles of designing such systems and introduce creative concepts and perspectives. These advantages will pave the way towards taking advantage of the computing power of future many-core systems.Broader Impacts: This BRIGE project aims to tightly integrate the research with different levels of education to attract traditionally underrepresented ethnic and gender groups into the engineering and computing pipeline and to pursue a successful higher education. An associated broadening participation plan will support a number of mentoring and outreach activities for disseminating knowledge of the advancements in modern microprocessor design into high school and university classrooms. It will include an outreach to K-12 students and teachers and local community college in a predominantly underprivileged and social-economically impacted neighborhood of Miami, Florida, leveraging upon the existing university programs to attract underrepresented students to the discipline of engineering. Several undergraduate researchers selected from a diversified student body will be actively involved with the proposed research activities. The integration of experimental effort along with theoretical analysis will offer invaluable experiences to graduate and undergraduate researchers involved in the project. The educational component also includes development of a specific course on many-core architecture and its interaction with operating systems, integrated with the research program. The advances expected from this project will enable the design of next generation high-performance microprocessors that will impact the computing ability and performance of future personal computers and the public at large.

智力优势：BRIGE研究项目的目标是利用当前的微处理器设计来支持和满足对系统性能/吞吐量日益增长的需求。微处理器复杂性的不断增加使得更新现有技术成为必要。该项目将探索众核微处理器的设计空间，重点是通过硬件支持减少操作系统开销和对用户应用程序的干扰。该项目预计将产生一个更详细的了解硬件和系统软件之间的相互作用，在执行应用程序的一部分。提出了一种将操作系统服务分布到专用核上的变革方案。这项研究的新奇在于试图将联合收割机在微架构设计的不同层次的研究前沿结合到一个统一的框架中，特别关注可扩展性、自适应性和改进的故障处理。该项目将促进对设计此类系统的原则的基本理解，并介绍创造性的概念和观点。这些优势将为利用未来众核系统的计算能力铺平道路。更广泛的影响：BRIGE项目旨在将研究与不同层次的教育紧密结合，以吸引传统上代表性不足的种族和性别群体进入工程和计算管道，并追求成功的高等教育。一个相关的扩大参与计划将支持一些辅导和推广活动，以便向高中和大学课堂传播现代微处理器设计进步的知识。它将包括在佛罗里达迈阿密的一个主要贫困和社会经济影响的社区，对K-12学生和教师以及当地社区学院的推广，利用现有的大学课程吸引代表性不足的学生到工程学科。从多元化的学生团体中选出的几名本科生研究人员将积极参与拟议的研究活动。实验工作的整合沿着理论分析将为参与该项目的研究生和本科生研究人员提供宝贵的经验。教育部分还包括开发关于众核架构及其与操作系统交互的特定课程，并与研究计划相结合。预计该项目的进展将使下一代高性能微处理器的设计成为可能，这将影响未来个人计算机和公众的计算能力和性能。