CI-New: Community Infrastructure Supporting Hardware Acceleration Research and Education

CI-New：支持硬件加速研究和教育的社区基础设施

• 批准号: 1730606
• 负责人: Kevin Skadron
• 金额: $ 49.96万
• 依托单位: University of Virginia Main Campus
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2022-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1730606&HistoricalAwards=false
• 关键词: CI; New; Community; Infrastructure; Supporting

Growth in data analytics is driving rapid development in specialized processor technologies, in the form of hardware accelerators. In addition to known accelerators such as the GPUs and FPGAs, exciting new architectures are offering capabilities not previously possible. To support the computer and information science and engineering (CISE) research community in understanding and making the most of these technologies, the University of Virginia (UVA) is developing a CISE Community Research Infrastructure to Support Hardware Acceleration Research and Education (UVA-SHARE).  This cyberinfrastructure gives system researchers access to an environment of diverse accelerators and tools to perform accelerator research, and provides application researchers the necessary resources to develop and test solutions that accelerate research computing applications. UVA-SHARE offers researchers a suite of clusters containing a diverse representation of the industry's latest accelerators.  UVA-SHARE users also have access to software tools, libraries, and technical support to make progress with research as soon as possible.  The broader impact of this project will primarily come in the form of the research it enables. Research enabled by UVA-SHARE is expected to make new advances in computer processor technologies, software infrastructure to support diverse accelerators, and enable new applications and new solutions in scientific computing.  The PI team will implement a set of community engagement activities to promote UVA-SHARE. Access and use of this cyberinfrastructure by early-career researchers is especially encouraged. The research will also achieve broader impact by training students in cutting-edge accelerator technologies and fluency across diverse technologies.

数据分析的增长正在以硬件加速器的形式推动专用处理器技术的快速发展。除了已知的加速器（例如GPU和FPGA）之外，令人兴奋的新架构还提供了以前不可能的功能。为了支持计算机和信息科学与工程（CISE）研究社区了解和充分利用这些技术，弗吉尼亚大学（UVA）正在开发CISE社区研究基础设施，以支持硬件加速研究和教育（UVA-SHARE）。 这种网络基础设施使系统研究人员能够访问各种加速器和工具的环境，以执行加速器研究，并为应用研究人员提供必要的资源，以开发和测试加速研究计算应用的解决方案。UVA-SHARE为研究人员提供一套集群，其中包含业界最新加速器的各种代表。UVA-SHARE用户还可以访问软件工具、库和技术支持，以便尽快取得研究进展。该项目的更广泛影响将主要体现在它所支持的研究方面。UVA-SHARE的研究有望在计算机处理器技术、支持各种加速器的软件基础设施方面取得新的进展，并在科学计算中实现新的应用和新的解决方案。PI团队将实施一系列社区参与活动，以促进UVA-SHARE。特别鼓励早期职业研究人员访问和使用这一网络基础设施。该研究还将通过培训学生掌握尖端加速器技术和熟练掌握各种技术来实现更广泛的影响。