CRI: Infrastructure acquisition for sub-100 nano VLSI research

摘要计划：NSF 04-588简明计算研究基础设施标题：CRI：100纳米以下VLSI研究的基础设施采购建议：CNS 0551621PI：Bhanja，Sanjukta研究所：南佛罗里达大学随着VLSI电路规模越来越小，进入纳米级范围，VLSI电路设计的挑战已经显著增加。该项目将获得计算、存储和软件资源，使研究人员能够设计工具来处理非常大的电路，并执行多目标、多层次的设计，以平衡功率、误差和热问题的要求。这些先进的计算资源将使团队能够探索数百万门的纳域电路的概率建模、电路的大型结构描述的设计和综合、用于大型设计验证的无刺激模型检查、基于神经网络的视频编码中的算法问题、模拟表征和设备性能模拟。该项目的更广泛影响将包括对研究生的直接培训以及综合和概率建模工具的传播。

AbstractProgram: NSF 04-588 CISE Computing Research InfrastructureTitle: CRI: Infrastructure acquisition for sub-100 nano VLSI researchProposal: CNS 0551621PI: Bhanja, SanjuktaInstitution: University of South Florida With increased scaling down of VLSI circuits, into the nano-scale range, the challenges of design of VLSI circuits has grown significantly. This project will acquire computational, storage and software resources to enable the investigators to design tools to handle very large circuits and to perform multi-objective, multi-level designs, to balance across requirements of power, error, and thermal issues. These advanced computational resources will enable the team to explore probabilistic modeling of nano-domain circuits for millions of gates, design and synthesis of large structural descriptions of circuits, stimulus-free model checking for verification of large designs, algorithmic issues in neural network based video coding, analog characterization, and device performance simulation. Broader impacts of the project will include direct training of graduate students and dissemination of tools for synthesis and probabilistic modeling.