Integrating Numerical Software Libraries with an InteractiveScientific Computing Evironment

将数值软件库与交互式科学计算环境集成

• 批准号: 9116710
• 负责人: Andrew Sherman
• 金额: $ 24.95万
• 依托单位: Scientific Computing Associates Inc
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-08-01 至 1995-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9116710&HistoricalAwards=false
• 关键词: Integrating; Numerical; Software; Libraries; InteractiveScientific

In terms of hardware capabilities, distributed/parallel computing is a technical reality today. With modern workstations and networking, users can compute on workstations on remote compute servers, or even on ensembles of computers, running compute-intensive kernels in parallel by exploiting idle CPU cycles on networks. However, substantial knowledge about networking and parallel/distributed computing is required to achieve this. The goal of this SBIR Phase II project is to build a state-of-the-art software package that will bring this immense potential power to the user's own workstation, except they will run many times faster. The package will be based on CLAM(R), a powerful, interactive software environment for computation and visualization, designed, developed and marketed by Scientific Computing Associates, Inc. The Phase I work proved the feasibility of seamlessly accessing local external libraries and demonstrated CLAMshell(TM), a Mac-like user- friendly front end for CLAM. The phase II work will complete the research on two approaches: (1) parallelizing the compute intensive kernals of CLAM's internal back end (2) enhancing CLAM's foreign procedure interface to permit access to procedures which run on other computers (including parallel supercomputers) or over a computer network in a distributed/parallel fashion.

就硬件能力而言，分布式/并行计算是一种 今天的技术现实 有了现代化的工作站和网络， 用户可以在远程计算服务器上的工作站上进行计算， 在计算机上运行计算密集型内核， 通过利用网络上的空闲CPU周期来并行。 然而，在这方面， 具备丰富的网络和并行/分布式知识 需要计算来实现这一点。 SBIR第二阶段项目的目标是建立一个最先进的 软件包，将带来这一巨大的潜力， 用户自己的工作站，除了他们将运行快很多倍。 的 软件包将基于CLAM（R），一个功能强大的交互式软件 环境的计算和可视化，设计，开发和 由Scientific Computing Associates，Inc. 第一阶段的工作证明了无缝访问本地 外部库，并演示了CLAMshell（TM），一个类似Mac的用户- 友好的前端蛤蜊。 第二阶段的工作将完成 研究了两种方法：（1）计算密集型并行化 CLAM内部后端的内核（2）增强CLAM的外部 过程接口，允许访问运行在其他 计算机（包括并行超级计算机）或通过计算机 以分布式/并行方式进行网络连接。