STTR Phase I: Sub-circuit Energy Monitoring and Performance ReAction Software (SEMPRA Software)

这个项目的更广泛的影响/商业潜力集中在提高计算机功率利用的效率上，即使晶体管的尺寸继续缩小。现代计算设备的能量密度要求比以往任何时候都要高，这导致移动设备的功率限制严格，固定基础设施计算的运营成本高昂，最高密度微芯片的产量低得令人难以置信。移动设备制造商在电池寿命方面展开了激烈的竞争。云服务提供商消耗的能源相当于小城市的能源，并为此买单。与此同时，美国制造商的竞争力悬而未决。本项目将通过在不增加功率的情况下提高IC性能，对微芯片、电子设备和智能电力需求-响应管理系统中的能源和电力管理产生广泛的影响。开启了计算机电力利用的新纪元。这个小型企业技术转移(STTR)第一阶段项目扩展了关于集成芯片的现场功率测量和功率管理的行业和学术研究。在很大程度上，到目前为止，工业和学术研究都缺乏现场功率测量的实用解决方案。然而，这个项目建立在之前由NSF资助的研究的基础上，这些研究指向了新的实际应用。通过消除工业级试验台中的重大技术风险，该STTR项目将展示其中一个应用的可行性，它将提供精确子组件级别的电源和温度监控、全封装芯片上的运行时热点检测以及基于准确电源/温度测量的智能电源管理。在这个项目完成之前，没有人能预测是否有可能以低到足以证明商业采用的管理费用来运营。

The broader impact/commercial potential of this project centers on increasing the efficiency of computer power utilization even as transistors continue to shrink in size. The energy density requirements of modern computing devices are higher than ever before, leading to strict power constraints in mobile devices, high operating costs for fixed infrastructure computing, and incredibly low yields for the highest-density microchips. Mobile device manufacturers fiercely compete on battery life. Cloud service providers consume, and pay for, the energy equivalent of small cities. Meanwhile, the competitiveness of American manufacturers hangs in the balance. The present project will have wide ranging impacts in Energy and Power Management in microchips, electronic devices, and smart power demand-response management systems that utilize them, by increasing IC performance without increasing power ? opening the door to a new era of computer power utilization. This Small Business Technology Transfer (STTR) Phase I project expands upon both industry and academic research concerning in-situ power measurement and power management of integrated chips. For the most part, to date both industry and academic research has produced a striking paucity of practical solutions for in-situ power measurement. This project, however, builds upon prior NSF-funded research which has pointed toward novel practical applications. By removing significant technical risks in an industry-class testbed, this STTR project will demonstrate the viability of one of these applications, which will provide accurate sub-component level power and temperature monitoring, run-time hotspot detection on a fully packaged chip, and intelligent power management based on accurate power/temperature measurements. Until this is project is complete, no one can predict whether it is possible to operate with overheads that are low enough to justify commercial adoption.