ITR: Collaborative Hardware-Software Adaptation for Multimedia Applications

ITR：多媒体应用的软硬件协同适配

• 批准号: 0205638
• 负责人: Sarita Adve
• 金额: $ 100万
• 依托单位: University of Illinois at Urbana-Champaign
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2002
• 资助国家: 美国
• 起止时间: 2002-10-01 至 2006-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0205638&HistoricalAwards=false
• 关键词: ITR; Collaborative; Hardware; Software; Adaptation

Mobile systems primarily processing multimedia data are expected to becomea dominant computing platform for many application domains. The design ofsuch systems imposes several new challenges, as it must consider demanding,dynamic, and multidimensional resource requirements and constraints, withenergy becoming a first-class resource. At the same time, the ability ofmultimedia applications to trade off output quality for system resourcesand the difference between their peak and average demands offers a hugeopportunity for optimization.A promising approach to meet the challenges of mobile multimedia systems,therefore, is to design all system layers with an ability to adapt inresponse to system or application changes. Further, to reap the fullbenefits of these adaptations, all system layers must cooperate to reach asystem-wide globally-optimal configuration. This research seeks to developand demonstrate an integrated cross-layer adaptive system where hardwareand all software layers cooperatively adapt to changing system resourcesand application demands, seeking to maximize user satisfaction whilemeeting resource constraints of energy, time, and bandwidth. This work isexpected to have a large impact because it will expose sources ofsubstantial performance improvement not available before, for a platform ofincreasing importance to many application domains.

主要处理多媒体数据的移动系统预计将成为许多应用领域的主导计算平台。 此类系统的设计提出了一些新的挑战，因为它必须考虑苛刻的、动态的、多维的资源需求和约束，而能源将成为一流的资源。 同时，多媒体应用能够权衡输出质量和系统资源，以及峰值和平均需求之间的差异，这为优化提供了巨大的机会。因此，应对移动多媒体系统挑战的一个有前途的方法是设计所有系统能够适应系统或应用程序变化的层。此外，为了获得这些适应的全部好处，所有系统层必须合作以达到系统范围内的全局最优配置。 本研究旨在开发和演示一种集成的跨层自适应系统，其中硬件和所有软件层协同适应不断变化的系统资源和应用需求，力求最大限度地提高用户满意度，同时满足能源、时间和带宽的资源限制。 这项工作预计会产生巨大的影响，因为对于一个对许多应用程序领域越来越重要的平台来说，它将揭示以前无法实现的实质性性能改进的来源。