Parallel Program Modeling and Evaluation

并行程序建模和评估

• 批准号: 8802283
• 负责人: Gary Nutt
• 金额: $ 16.74万
• 依托单位: University of Colorado at Boulder
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1988
• 资助国家: 美国
• 起止时间: 1988-08-15 至 1991-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8802283&HistoricalAwards=false
• 关键词: Parallel; Program; Modeling; Evaluation

Developing large programs for parallel, distributed computer systems is at least as difficult as developing software for more traditional sequential systems. The added difficulty is in the innovation required to exploit the multiplicity of processors. Considerable effort has been expended to build operating and networks so that they are conducive to distributed programming, and new programming languages have been built to increase the expressive computation so that it takes advantage of the structure of the support system. Program designers do not have adequate tools to assist them in creating and evaluating parallel, distributed computations. They also lack convenient ways to describe, model and visualize these computations. The goal of the research is to study mechanism that support the orderly exploration of parallel program structures, and which provide varying amounts of feedback evaluation during that design process.

开发用于并行的分布式计算机系统的大型程序至少与为更传统的顺序系统开发软件一样困难。 额外的困难在于利用处理器多样性所需的创新。 已经花费了巨大的努力来构建操作和网络，以便它们有利于分布式编程，并且已经构建了新的编程语言来增加表现力的计算，从而利用了支持系统的结构。 程序设计师没有足够的工具来帮助他们创建和评估并行的分布式计算。 他们还缺乏描述，建模和可视化这些计算的方便方法。 该研究的目的是研究支持平行程序结构有序探索的机制，并在该设计过程中提供不同数量的反馈评估。