Collaborative Research: Frameworks: Advancing Computer Hardware and Systems' Research Capability, Reproducibility, and Sustainability with the gem5 Simulator Ecosystem

协作研究：框架：利用 gem5 模拟器生态系统提升计算机硬件和系统的研究能力、可重复性和可持续性

• 批准号: 2311888
• 负责人: Jason Lowe-Power
• 金额: $ 242.5万
• 依托单位: University of California-Davis
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-10-01 至 2027-09-30
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2311888&HistoricalAwards=false
• 关键词: Collaborative; Research; Frameworks; Advancing; Computer

This project aims to improve the cyberinfrastructure ecosystem for computer system research by expanding and improving the popular gem5 simulation infrastructure. The breakdown in Moore’s Law and Dennard Scaling is leading to a drastic increase in the scale of computing systems, making it increasingly difficult to make broad scientific progress in research to improve computer systems without considering the entire scale of the hardware-software stack from transistors to applications. This project will build on prior NSF investments and scale the gem5 simulator to more users, more types of computing systems, larger and higher fidelity computing systems, and more scientific communities beyond computer architecture. These improvements to the gem5 simulation infrastructure will enable researchers to design and understand the next generation of supercomputers, laptops, and mobile devices. It also supports education and diversity by creating teaching materials to broaden the participation in computer systems research for both graduate and undergraduate students.Working with a large and diverse team of experts from the computer architecture community, this project will develop a wide variety of improvements to the gem5 simulation infrastructure. This includes modeling future devices and phenomena such as reliability, security, and chiplets; creating scalable models for modern hardware by improving the accuracy of current models, adding support for emerging vector extensions, and improving support for common accelerators such as GPUs; scaling simulation performance by optimizing gem5’s performance and providing interoperability with other simulation infrastructure; improving modularity, reusability, and reproducibility by making baseline system easier to use and providing ready-to-use benchmarks for many domains; and sustaining the community with outreach and education by running workshops, tutorials, bootcamps and expanding gem5’s userbase beyond computer architects through developing asynchronous learning and teaching tools. Additionally, it will help ensure the long-term sustainability of the gem5 infrastructure by growing a sustainable ecosystem and increasing participation in computer system research.This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Division of Computing and Communications Foundations within the Directorate for Computer and Information Science and Engineering.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

该项目旨在通过扩展和改进流行的 gem5 模拟基础设施来改善计算机系统研究的网络基础设施生态系统。摩尔定律和登纳德缩放比例的崩溃导致计算系统规模急剧增加，如果不考虑从晶体管到应用程序的软硬件堆栈的整个规模，就越来越难以在改进计算机系统的研究中取得广泛的科学进展。该项目将建立在 NSF 之前投资的基础上，并将 gem5 模拟器扩展到更多用户、更多类型的计算系统、更大和更高保真度的计算系统以及计算机体系结构之外的更多科学社区。 gem5 模拟基础设施的这些改进将使研究人员能够设计和了解下一代超级计算机、笔记本电脑和移动设备。它还通过创建教材来支持教育和多样性，以扩大研究生和本科生对计算机系统研究的参与。该项目将与来自计算机体系结构社区的庞大且多元化的专家团队合作，对 gem5 模拟基础设施进行各种改进。这包括对未来的设备和现象进行建模，例如可靠性、安全性和小芯片；通过提高当前模型的准确性、增加对新兴矢量扩展的支持以及改进对 GPU 等常见加速器的支持，为现代硬件创建可扩展模型；通过优化 gem5 的性能并提供与其他模拟基础设施的互操作性来扩展模拟性能；通过使基线系统更易于使用并为许多领域提供即用型基准来提高模块化、可重用性和可重复性；通过举办研讨会、教程、训练营来维持社区的外展和教育，并通过开发异步学习和教学工具将 gem5 的用户群扩展到计算机架构师之外。此外，它将通过发展可持续的生态系统和增加对计算机系统研究的参与来帮助确保gem5基础设施的长期可持续性。该奖项由高级网络基础设施办公室颁发，并得到计算机和信息科学与工程理事会内的计算和通信基金会部的共同支持。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值进行评估，被认为值得支持 以及更广泛的影响审查标准。