New Technologies for Very Large-Scale Electronic Design Automation

超大规模电子设计自动化新技术

• 批准号: RGPIN-2014-04098
• 负责人: Rakai, Logan
• 金额: $ 1.6万
• 依托单位: University of Calgary
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=637179
• 关键词: New; Technologies; Very; Large; Scale

The field of electronic design automation is concerned with creating software capable of automating the process of designing electronic devices. The electronics industry is known for exponential increases in design complexity and performance of resulting circuits [1]. At the same time, new technologies are being developed in other fields to solve difficult problems, for example, machine learning, MapReduce programming models, and high performance computing systems. The research program outlined in what follows is aimed at developing new ways to solve very large scale problems in electronic design automation using new technologies. Major challenges being faced in the field are related to problems arising from scale. For example, the very large-scale problems arising out of creating circuits with billions of components. Industry is constantly feeling pressure to quickly create devices for competitive producers of electronics including computers, phones, and tablets. Automation is the key to being able to design quickly, yet the most challenging designs require substantial intervention from designers to configure the software for each design. Many iterations of a design can take place before converging to a final design meeting product specifications. The design process may take between several months to over a year for cutting-edge circuits. Hence, the targets may not be met in time to meet the producer’s deadlines. Failing to meet a deadline can equate to big losses and loss of future customers in business. These factors all go to support research into solving large problems at scales never previously seen before.The ultimate goal of the research program is to develop several pieces which will be the beginning of a highly scalable design automation tool. The tool will be available through a cloud service where the number of computing cores is elastic and practically limitless. Consumers would be able to have designs of arbitrary size and pay for more or less computing resources depending on how quickly they need the circuit fabricated. Having such a tool would greatly benefit industry and the projects described in this proposal will bring this vision closer to reality and inspire other researchers to pursue a similar goal.The projects discussed in the proposal have a direct impact on the multi-billion dollar electronic design industry. However, the research also involves topics that are of strategic importance to Canada. This is evidenced by the overlap with NSERC's strategic grants target areas in information and communication technologies that emphasize a close connection with industry and recognize market potential of research, collaboration, technical and financial implications, and realizing value of the research within five years. In particular, the categories of Next Generation Computing Platforms: cloud computing, From Data to Knowledge to Action: unstructured pattern recognition, and advanced applications in artificial intelligence. The proposed research program also has the support of the department and aligns with the University of Calgary’s strategic research plans to become a top five Canadian research university by 2016. The proposed research program aligns with the Eyes High strategic vision in the categories of balance, curiosity, collaboration, communication, globalization, and excellence. Taken together, they support programs invested in new lines of inquiry with a global impact and encourage interdisciplinary collaboration.

电子设计自动化领域关注的是创建能够使电子设备设计过程自动化的软件。电子工业以设计复杂性和电路性能的指数增长而闻名。与此同时，其他领域正在开发新技术来解决难题，例如机器学习、MapReduce编程模型和高性能计算系统。以下概述的研究计划旨在开发使用新技术解决电子设计自动化中大规模问题的新方法。该领域面临的主要挑战与规模问题有关。例如，由数十亿个元件组成的电路所产生的大规模问题。工业界不断感到压力，必须迅速为有竞争力的电子产品生产商（包括电脑、电话和平板电脑）制造设备。自动化是能够快速设计的关键，然而最具挑战性的设计需要设计师的大量干预来配置每个设计的软件。在最终设计符合产品规格之前，可能会进行多次设计迭代。尖端电路的设计过程可能需要几个月到一年多的时间。因此，目标可能无法及时满足生产商的最后期限。不能在最后期限前完成任务就意味着巨大的损失和未来客户的流失。这些因素都支持研究以前所未有的规模解决大问题。研究计划的最终目标是开发几个部分，这将是一个高度可扩展的设计自动化工具的开始。该工具将通过云服务提供，其中计算核心的数量是弹性的，实际上是无限的。消费者将能够拥有任意尺寸的设计，并根据他们需要制造电路的速度来支付更多或更少的计算资源。拥有这样的工具将极大地有利于工业，本提案中描述的项目将使这一愿景更接近现实，并激励其他研究人员追求类似的目标。建议中讨论的项目对价值数十亿美元的电子设计行业有直接影响。然而，该研究也涉及对加拿大具有战略重要性的主题。这一点可以从NSERC在信息和通信技术领域的战略资助重叠中得到证明，这些领域强调与行业的密切联系，并认识到研究、合作、技术和财务影响的市场潜力，并在五年内实现研究的价值。特别是下一代计算平台的类别：云计算，从数据到知识到行动：非结构化模式识别，以及人工智能的高级应用。拟议的研究项目也得到了该部门的支持，并与卡尔加里大学的战略研究计划保持一致，该计划到2016年成为加拿大排名前五的研究型大学。拟议的研究项目与Eyes High在平衡、好奇心、协作、沟通、全球化和卓越等方面的战略愿景一致。总之，它们支持投资于具有全球影响力的新研究领域的项目，并鼓励跨学科合作。