NER: Y-Junction Nanotube-based Computer Devices and Architectures

项目编号：304284主要研究人员：R. Iris Bahar, Jie Chen， Joseph mundist：布朗大学项目名称：NER：基于y结纳米管的计算机设备和体系结构本研究侧重于纳米尺度计算的探索性研究。与传统的硅基计算相比，这些新的计算形式将允许更有效地利用纳米级电子和设备。特别是，本研究的重点是寻找一种有效的方法来处理纳米级器件中发现的高缺陷率，并开发一种动态缺陷容限的体系结构。目前正在开发三个领域：(1)基于y结碳纳米管（CNTs）的器件模型正在被定义。这项工作可能包括使用自组装能力来编程CNTdevices的互连和构建I/O接口。(2)提出了一种基于马尔可夫随机网络和短传播的基于概率的纳米级计算机设备设计方法。这种方法允许架构将错误作为概率最大化的自然结果来适应，从而消除了实际检测故障的需要。2003年秋季将开设一门纳米计算的新课程。本课程将提供一个独特的机会，让学生接触到影响计算机的广泛学科。

Proposal Number: 304284Principal Investigators: R. Iris Bahar, Jie Chen, Joseph MundyInst: Brown UniversityProposal Title: NER: Y-Junction Nanotube-based Computer Devices and ArchitecturesThis research focuses on exploratory research in nano-scale computation. These new forms of computation will allow for moreefficient means of exploiting nano-scale electronics and devices thanthose based on traditional silicon-based computation. In particular,this research focuses on finding an effective means of dealing withhigh defect rates found in nano-scale devices and developing anarchitecture that is dynamically defect tolerant.Three areas are being developed: (1) Device models are being definedbased on Y-junction carbon nanotubes (CNTs). This work may includeuse of self-assembly capabilities to program interconnections of CNTdevices and build I/O interfaces. (2) A probabilistic-based designmethodology for nano-scale computer devices will be proposed based onthe Markov random network and brief propagation. This methodologyallows for an architecture that adapts to errors as a naturalconsequence of probability maximization, thereby removing the need toactually detect faults. (3) A new course in nano-computing will beintroduced in fall 2003. The course will offer a unique opportunityto expose students to the wide range of disciplines that influencenano-computing.