课题基金 / 基金详情

CAREER: Predicting Timing Violations: A New Direction for Robust System Design

CAREER: Predicting Timing Violations: A New Direction for Robust System Design
职业:预测时序违规:鲁棒系统设计的新方向
批准号:
1253024
负责人:
Sanghamitra Roy
金额:
$46.61万
依托单位:
依托单位国家:
美国
项目类别:
Continuing Grant
财政年份:
2013
资助国家:
美国
项目状态:
已结题
起止时间:
2013-03-01 至 2021-02-28

项目摘要

项目成果

Sanghamitra Roy的其他基金

相似基金

相关文献

中文摘要
翻译
时序违规是快速技术扩展的产物,体现了微处理器设计中的核心可靠性挑战。在时间违规方面,现有的大量技术可分为两大类:被动和主动。在这两种方案中,缺乏足够的时间来识别即将到来的时间冲突限制了纠正技术的范围,分别导致了巨大的恢复开销或故障覆盖损失。通过结合应用程序、体系结构和电路层的信息进行跨层分析,这个CAREER项目证明了可以提前几个时钟周期预测时间冲突。时间违逆的早期预测可以为稳健的系统设计提供巨大的杠杆作用,为未来的系统研究开辟了一个有希望的方向。在这种范例中,微处理器可以以更紧凑的频率工作,在这种情况下,可预测的错误经常发生,并且可以容忍最小的性能损失。这样的系统设计风格可以重塑物理设计算法,推动微处理器组件的节能前沿。跨层设计这一新兴领域的研究将使电路设计师和系统架构师能够加强协同设计,并开发出价格合理、节能可靠的计算机系统。集成电路的快速增长和发展创造了计算机工程技能的巨大需求,以应对即将到来的挑战。CAREER项目旨在通过集中知识传播,培养一批能够解决集成电路日益增长的不可靠性问题的未来工程师。将通过犹他州立大学现有的平台,如女工程师协会(SWE)、工程州和妇女与性别中心(CWG)项目,积极寻求本科生、女性和少数族裔的参与。
英文摘要
Timing violations, an artifact of rapid technology scaling, embody a central reliability challenge in microprocessor design. A vast body of existing techniques in the landscape of timing violations fall under two broad categories: reactive and proactive. In both of these schemes, the lack of sufficient time in identifying an upcoming timing violation limits the scope of corrective techniques, incurring a large recovery overhead, or loss in fault coverage, respectively. Using a cross-layer analysis combining information from the application, architecture and circuit layers, this CAREER project demonstrates that timing violations can be predicted several clock cycles early. Early prediction of timing violations can offer a vast leverage in robust system design, opening up a promising direction for future systems research. In this paradigm, microprocessors can operate at a tighter frequency, where predictable errors frequently occur and are tolerated with minimal performance loss. Such a system design style can reshape physical design algorithms, boosting the energy efficient frontiers for microprocessor components. Research in the nascent area of cross-layer design will enable circuit designers and system architects to increase collaborative design, and develop affordable, energy efficient and reliable computer systems. The rapid growth and evolution of integrated circuits creates a huge demand of Computer Engineering skill sets for addressing upcoming challenges. The CAREER project seeks to create a stream of future engineers capable of tackling growing unreliability in integrated circuits through focused knowledge dissemination. Undergraduates, women and minorities will be actively sought for participation through the existing platforms at Utah State University such as the Society of Women Engineers (SWE), Engineering State and Center for Women and Gender (CWG) programs.
期刊论文(1)
专著(0)
科研奖励(0)
会议论文
DOI: 10.1109/dac18074.2021.9586224
发表时间: 2021-12
期刊: 2021 58th ACM/IEEE Design Automation Conference (DAC)
影响因子: --
作者: [Pramesh Pandey;N. D. Gundi;Koushik Chakraborty;Sanghamitra Roy]
通讯作者: Pramesh Pandey;N. D. Gundi;Koushik Chakraborty;Sanghamitra Roy
CNS Core: Small: Ultra Low Power Hardware AI Accelerator for Training at the Edge
  • 批准号:
    2106237
  • 项目类别:
    Standard Grant
  • 资助金额:
    $49.93万
  • 财政年份:
    2021
  • 负责人:
    Sanghamitra Roy
  • 依托单位:
CSR: Small: DARP: Promoting Energy Efficient System Design Through a Dynamically Adaptable Resilient Pipeline
  • 批准号:
    1421022
  • 项目类别:
    Standard Grant
  • 资助金额:
    $44.17万
  • 财政年份:
    2014
  • 负责人:
    Sanghamitra Roy
  • 依托单位:
CSR:Small:Employing Design Automation to Build Foundations for Holistic Multicore Design
  • 批准号:
    1117425
  • 项目类别:
    Standard Grant
  • 资助金额:
    $37.0万
  • 财政年份:
    2011
  • 负责人:
    Sanghamitra Roy
  • 依托单位:
海外基金