Paradigms for Parallel Algorithms

并行算法范式

• 批准号: 9111348
• 负责人: Uzi Vishkin
• 金额: $ 25.5万
• 依托单位: University of Maryland, College Park
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-01-01 至 1996-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9111348&HistoricalAwards=false
• 关键词: Paradigms; Parallel; Algorithms

The knowhow of designing efficient algorithms meets the definition of a technology. This proposal is about various aspects of a parallel algorithmic technology. Given two parallel algorithms for the same problem one is more efficient than the other if: (1) primarily, its time-processor product is smaller, and (2) secondarily (but important), its parallel time is smaller. It is a primary intellectual challenge to get the fastest possible time, by a processor-efficient algorithm. The proposed work identifies several considerable challenges on which to work in the next few years. Contributions to major application domains ("Grand Challenges"), such as computer vision and comparison of molecular sequences, and deadline-driven computing, as well as to the fundamentals of the evolving technology of parallel algorithms are expected.

设计高效算法的专有知识符合技术的定义。本提案是关于并行算法技术的各个方面。对于同一个问题，给定两个并行算法，其中一个比另一个更有效，如果：(1)首先，它的时间处理器乘积更小，(2)其次（但很重要），它的并行时间更小。通过处理器效率高的算法获得尽可能快的时间是一个主要的智力挑战。拟议的工作确定了今后几年需要应对的若干重大挑战。期望对主要应用领域（“大挑战”）做出贡献，例如计算机视觉和分子序列的比较，以及截止日期驱动的计算，以及并行算法发展技术的基础。