Cross-Layer Intelligent System-Based Adaptive Power Conditioning for Robust and Reliable Mixed-Signal Multi-Core SoCs

基于跨层智能系统的自适应功率调节，用于稳健可靠的混合信号多核 SoC

• 批准号: 1255754
• 负责人: Yiorgos Makris
• 金额: $ 19.2万
• 依托单位: University of Texas at Dallas
• 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
• 起止时间: 2013-04-01 至 2016-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1255754&HistoricalAwards=false
• 关键词: Cross; Layer; Intelligent; System; Based

This project seeks to investigate the effectiveness of adaptive power conditioning in averting failures and enhancing reliability of mixed-signal multi-core Systems-on-Chips (SoCs), while retaining power efficiency. The investigation aims to leverage the increasing integration of non-intrusive sensors for monitoring various operational parameters and cost-effective tuning knobs for calibrating circuit performance through online power condition adjustment. The project aims to develop a cross-layer intelligent system, which will continuously adjust the flow of the entire power distribution network based on the feedback provided by the sensors. Ultimately, the aspiration of this project is to demonstrate that a holistic approach to adaptive power conditioning can (i) ensure operational robustness while minimizing power consumption, and (ii) enhance reliability and prolong life expectancy of integrated circuits. The intellectual contribution of this work will be in the development of a practical, yet theoretically rooted and analytically supported intelligent system for capturing correlations between circuit performance, sensor readings, and calibration knob positions, and utilizing them to improve robustness and reliability. The proposed research is complemented by educational and outreach activities demonstrating the principle investigators' (PIs') commitment to an integrated role as researchers, educators, and active community members. A new graduate-level course will be developed and offered by the PIs, who have proven track records of involving the students of various levels in their research activities, including women and other under-represented groups. Several opportunities for participation in interdisciplinary research spanning Electrical Engineering, Computer Science, and Applied Mathematics as well as internships with industrial partners are expected.

该项目旨在调查自适应功率调节在避免故障中的有效性，并提高混合信号多核系统对芯片（SOC）的可靠性，同时保持功率效率。该调查旨在利用非侵入性传感器的增加集成，以监视各种操作参数和成本效益的调整旋钮，以通过在线功率条件调整来校准电路性能。该项目旨在开发跨层智能系统，该系统将根据传感器提供的反馈不断调整整个电源分配网络的流动。最终，该项目的愿望是证明一种自适应功率调节的整体方法可以（i）确保运行鲁棒性同时最大程度地减少功耗，以及（ii）提高综合电路的可靠性和延长的预期寿命。这项工作的智力贡献将是开发实用的，但理论上根深蒂固的和分析支持的智能系统，以捕获电路性能，传感器读数和校准旋钮位置之间的相关性，并利用它们来提高鲁棒性和可靠性。拟议的研究得到了教育和外展活动的补充，这些活动证明了主要研究人员（PIS'）对研究人员，教育工作者和活跃社区成员的综合作用的承诺。 PIS将开发和提供新的研究生课程，他们已证明与各个级别的学生一起参与其研究活动（包括妇女和其他代表性不足的群体）的记录。期望参加跨学科研究的一些机会，这些研究涵盖电气工程，计算机科学和应用数学以及与工业合作伙伴的实习机会。