U.S.-Korea Cooperative Research: Modeling and Analyzing Communication Behavior of Parallel Programs on Distributed-Memory Multiprocessors

美韩合作研究：分布式内存多处理器上并行程序的通信行为建模与分析

• 批准号: 9722545
• 负责人: Andrew Sohn
• 金额: $ 2.43万
• 依托单位: New Jersey Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-08-15 至 2000-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9722545&HistoricalAwards=false
• 关键词: U.S.; Korea; Cooperative; Research; Modeling

9722545 Sohn This award supports a two-year collaborative research project between Dr. Andrew Sohn of the New Jersey Institute of Technology and Dr. Dongseung Kim of Korea University in Seoul, Korea on "Modeling and Analyzing Communication Behavior of Parallel Programs on Distributed-Memory Multiprocessors." The objective of the project are to 1) model the capabilities of hardware and software for overlapping communication with computation, and 2) compare the overlapping capabilities of distributed-memory multiprocessors on selected benchmark problems. Distributed-memory multiprocessors are known to have the architecture needed for increasing the machine's capability by simply adding more processor to the system. Programming distributed-memory machines, however, is often difficult due to many practical issues including algorithm design, data distribution, and workload distribution. When a large number of processors work together to solve a single problem, communication between processors becomes key to determining the performance of large scale distributed memory machines. This project will combine the expertise and facilities of the U.S. and Korea research groups in addressing this important problem in computer science. ***

小行星9722545 该奖项支持新泽西理工学院的Andrew Sohn博士和韩国首尔高丽大学的Dongseung Kim博士之间为期两年的合作研究项目，该项目的主题是"分布式内存多处理器上并行程序的通信行为建模和分析"。该项目的目标是：1）模拟硬件和软件的重叠通信与计算的能力，2）比较分布式内存多处理器在选定的基准问题上的重叠能力。 分布式存储器多处理器具有通过简单地向系统添加更多处理器来增加机器能力所需的体系结构。 然而，由于算法设计、数据分布和工作负载分布等许多实际问题，对分布式内存机器进行编程通常很困难。 当大量处理器协同工作以解决单个问题时，处理器之间的通信成为决定大规模分布式存储器机器性能的关键。 这个项目将联合收割机的专业知识和设施的美国和韩国的研究小组在解决这个重要的问题，在计算机科学。 ***