CI-NEW: Collaborative Research: Developing a Community Infrastructure for Reliability-Aware Cross-Layered Design of Integrated Circuits

CI-NEW：协作研究：为集成电路的可靠性感知跨层设计开发社区基础设施

• 批准号: 1629853
• 负责人: William Robinson
• 金额: $ 15万
• 依托单位: Vanderbilt University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-09-01 至 2020-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1629853&HistoricalAwards=false
• 关键词: CI; NEW; Collaborative; Research; Developing

Reliability analysis of integrated circuits (ICs) has become a greater concern in the development of computing systems. Technology scaling offers more transistors to enable system-on-a-chip (SoC) integration, but it also means more transistors that can be affected by radiation-induced faults. Designers must consider how those faults impact the entire computing system by including levels of abstraction in both hardware and software. Reliability cannot be addressed by considering the device, circuit, architecture, or application layer individually. Cross-layered modeling and simulation enables a better understanding of the complex interactions among particle physics, fabrication technology, processor architecture, and the software stack in order to develop reliable computing systems. With the increase in cost and time of the system design process, pre-fabrication modeling of radiation effects has become a critical field of research. By understanding these processes, designers can meet traditional constraints of performance and power, while addressing concerns for the soft error rate (SER). Potential application domains include, but are not limited to, automotive embedded systems, high-performance computing, and spaceborne environments. This project creates a research infrastructure that links soft error simulations at the device, circuit, architecture, and application levels. This infrastructure provides a testbed for characterizing the resilience of algorithms for computer and computational sciences. It also contains testbeds for evaluating the reliability of the architecture and the design of computing systems. By providing up-to-date models of the effects of single events on current and future technologies, it facilitates the development of computing models and technologies based on emerging scientific ideas from the research community. The infrastructure will be accessible to a larger research community through web-based content (e.g., online tutorials and webinars) and interactive workshops at conferences and symposia. This project broadens the participation of groups traditionally underrepresented in the areas of science, technology, engineering, and mathematics (STEM) through outreach to encourage students to enter the science and engineering fields. The STEM Scholars Program at North Carolina A&T provides a computing skills workshop to African-American high school students.

综合电路（IC）的可靠性分析已成为计算系统开发的更加关注。技术扩展提供了更多的晶体管以实现芯片上的系统（SOC）集成，但这也意味着更多的晶体管可能会受到辐射诱导的断层的影响。设计人员必须考虑这些故障如何通过在硬件和软件中包括抽象水平来影响整个计算系统。可靠性无法通过单独考虑设备，电路，体系结构或应用程序层来解决。跨层建模和仿真可以更好地了解粒子物理，制造技术，处理器体系结构和软件堆栈之间的复杂相互作用，以开发可靠的计算系统。随着系统设计过程的成本和时间的增加，辐射效应的预制前建模已成为研究的关键领域。通过了解这些过程，设计师可以达到传统的性能和权力限制，同时解决了对软错误率（SER）的担忧。潜在的应用领域包括但不限于汽车嵌入式系统，高性能计算和太空传播环境。该项目创建了一个研究基础架构，该研究基础架构可以在设备，电路，体系结构和应用程序级别上链接软错误模拟。该基础架构提供了一个测试床，以表征计算机和计算科学算法的弹性。它还包含用于评估架构可靠性和计算系统设计的测试床。通过提供单个事件对当前和未来技术的影响的最新模型，它促进了基于研究界新兴的科学思想的计算模型和技术的开发。大型研究社区将通过基于Web的内容（例如在线教程和网络研讨会）以及会议和研讨会上的交互式研讨会来访问基础架构。该项目扩大了传统上代表性不足在科学，技术，工程和数学（STEM）领域的团体的参与，以鼓励学生进入科学和工程领域。北卡罗来纳州A＆T的STEM学者计划为非裔美国高中生提供了计算技能研讨会。