EAGER: Collaborative Research: Heterogeneous Cores, Memory-Hierarchy and Communication Architectures for Future CMPs

EAGER：协作研究：未来 CMP 的异构核心、内存层次结构和通信架构

• 批准号: 1147406
• 负责人: Luca Carloni
• 金额: $ 10万
• 依托单位: Columbia University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-09-01 至 2013-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1147406&HistoricalAwards=false
• 关键词: EAGER; Collaborative; Research; Heterogeneous; Cores

Computing has enabled immense innovations in many fields and sociallife due to continued performance, power, and cost improvementsenabled by technology scaling. Unfortunately, two key trends threatenperformance and cost improvement of computers going into thefuture. First, technology scaling is at jeopardy, leading to majorchallenges in power consumption, reliability, and performance. Second,power and energy consumption has become a key constraint, yet existingsystems are mostly designed in a one-size-fits-all way agnostic to theneeds of different applications. To overcome both problems, thisresearch investigates novel uses of heterogeneous technologies inthree key components of a computing system: cores, interconnect, andmemory. The approach taken is an application-driven approach that aimsto seamlessly integrate heterogeneity in the three components. Majorexpected contributions of the research include: (1) an initial studyof first-principles based and application-driven design of cores, (2)new mechanisms for enabling phase-change memory based heterogeneousmain memory, (3) exploration of tradeoffs in the design of3-dimensional optical interconnects, (4) initial exploration of theinteraction of heterogeneous components in cores, interconnect, andmemory.The proposed research has the potential to transform the design andarchitecture of future multi-core systems, which are already a part ofthe entire IT sector and our daily lives. It can enable overcoming keychallenges that impediment higher-performance and lower-powerlower-cost computing, which has traditionally enabled new applicationsand discoveries. Enabling fundamentally efficient heterogeneousmulti-core systems can largely reduce energy and technology-scalingcosts of computing, and improve dependability and performance. Directtransfer of many ideas to industry are expected through extensivecollaborations with platform and chip design industries.

由于技术扩展带来的性能、功耗和成本的持续改进，计算在许多领域和社会生活中实现了巨大的创新。不幸的是，有两个主要趋势威胁着未来计算机的性能和成本改进。首先，技术扩展面临危险，导致功耗、可靠性和性能方面面临重大挑战。其次，电力和能源消耗已成为一个关键限制，但现有系统大多采用一刀切的方式设计，与不同应用的需求无关。为了克服这两个问题，本研究研究了异构技术在计算系统的三个关键组件中的新用途：核心、互连和内存。所采用的方法是应用程序驱动的方法，旨在无缝集成三个组件中的异构性。该研究的主要预期贡献包括：（1）基于第一原理和应用驱动的内核设计的初步研究，（2）实现基于相变存储器的异构主存储器的新机制，（3）探索设计中的权衡维光学互连，（4）对核心、互连和存储器中异构组件相互作用的初步探索。所提出的研究有可能改变未来多核系统的设计和架构，这些系统已经是整个 IT 行业和我们的日常生活。它可以克服阻碍更高性能和更低功耗、更低成本计算的关键挑战，而传统上这些挑战促进了新的应用和发现。 从根本上实现高效的异构多核系统可以大大降低计算的能源和技术扩展成本，并提高可靠性和性能。通过与平台和芯片设计行业的广泛合作，许多想法有望直接转移到行业。