EAGER: Data-Mining Driven Power-Efficient Intelligent Memory Storage for Mobile Video Applications

EAGER：适用于移动视频应用的数据挖掘驱动型节能智能内存存储

• 批准号: 1514780
• 负责人: Na Gong
• 金额: $ 20万
• 依托单位: North Dakota State University Fargo
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-07-01 至 2018-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1514780&HistoricalAwards=false
• 关键词: EAGER; Data; Mining; Driven; Power

Mobile devices such as smart-phones and tablets have become the most important medium for delivering Internet traffic, especially multimedia content, to end users. One of the most popular multimedia applications is video streaming. During this process, video decoding has become the dominant energy-intensive application used in mobile devices. In particular, the major signal processing units in video decoders, such as motion estimation and compensation, forward and inverse discrete cosine transform, require a significant amount of calculations and frequent embedded memory accesses. It is understood that embedded SRAM consumes a large amount of power and limits battery life, and this situation is only expected to grow with the emerging popularity of high quality mobile video applications. This project proposes to address this problem by incorporating advanced data mining techniques particularly suited to mobile video data applications into the hardware design process to yield an intelligent memory having high power efficiency. The PIs will explore and characterize the behaviors of video data and provide a better-informed low power hardware design. The goal is to create new power efficient mobile video memory designs that utilize the identified characteristics extracted by suitable data-mining techniques tailored to video data, which will serve as a core foundation to bring about drastic improvements in energy efficiency. The exploration of these intelligent low power techniques through the interaction of both hardware and software viewpoints will enable a new dimension for power savings. The success of this project will have a huge impact on the mobile computing community, architecture community, and everyday life. This project will also serve as an excellent educational platform to improve the understanding of green computing amongst future computer scientists and computer engineers. The PIs will jointly develop course modules focusing on software/hardware co-design for mobile devices, which can be integrated into a variety of different courses. The PIs will also continue to recruit underrepresented students, such as females and minorities, to participate in this project.

智能手机和平板电脑等移动设备已成为向最终用户提供互联网流量（尤其是多媒体内容）的最重要媒介。最流行的多媒体应用之一是视频流。在此过程中，视频解码已成为移动设备中使用的主要能源密集型应用。特别是视频解码器中的主要信号处理单元，例如运动估计和补偿、正向和反向离散余弦变换，需要大量的计算和频繁的嵌入式存储器访问。据了解，嵌入式SRAM会消耗大量电量并限制电池寿命，而且随着高质量移动视频应用的日益普及，这种情况预计只会进一步加剧。 该项目建议通过将特别适合移动视频数据应用的先进数据挖掘技术融入到硬件设计过程中来解决这个问题，以产生具有高功率效率的智能存储器。 PI 将探索和表征视频数据的行为，并提供更明智的低功耗硬件设计。目标是创建新的节能移动视频内存设计，利用针对视频数据定制的适当数据挖掘技术提取的识别特征，这将成为大幅提高能源效率的核心基础。通过硬件和软件观点的交互来探索这些智能低功耗技术将为节能开辟新的维度。该项目的成功将对移动计算社区、架构社区和日常生活产生巨大影响。 该项目还将作为一个优秀的教育平台，以提高未来计算机科学家和计算机工程师对绿色计算的理解。 PI将联合开发专注于移动设备软硬件协同设计的课程模块，这些模块可以集成到各种不同的课程中。 PI 还将继续招募女性和少数族裔等代表性不足的学生参与该项目。