CISE Research Instrumentation: Parallel and Distributed Constraint Programming Systems on Multiprocessor PCs: Implementations and Applications

9729848 Gupta，Gopal新墨西哥州立大学CISE仪器：多处理器PC上的并行和分布式约束编程系统：实施和应用该研究仪器用于支持购买基于奔腾的四处理器多处理器和一个快速以太网集线器，以支持NMSU计算机科学的下列并行和分布式计算项目：-并行和分布式约束逻辑编程系统，-用于互联网编程的基于并行约束的统一框架，-并行基于约束的自然语言处理，以及-并行规范和实时系统验证，所有这些都共享基于约束的技术的共同主题。对于所提到的项目，购买的设备有助于：第一，利用逻辑和约束程序中隐含的并行性的系统的设计和实施；第二，分布式网络使互联网编程的统一框架的原型成为可能，包括符号约束和并发性；第三，并行约束软件技术将显著加快大型自然语言处理应用程序的执行；第四，开发用于验证大型复杂实时系统的并行工具。并行性和分布性是实现这些项目目标的关键。所选择的平台具有成本效益、灵活性，并在共享内存多处理器和分布式处理器网络之间提供了平衡，前者效率较低但不可伸缩，后者效率较低但可伸缩。

9729848 Gupta, Gopal New Mexico State University CISE Instrumentation: Parallel and Distributed Constraint Programming Systems on Multiprocessor PCs: Implementations and Applications This research instrumentation serves to support the purchase of four-processor Pentium-based multiprocessors and a fast Ethernet hub to enable the following projects in parallel and distributed computing at the NMSU Computer Science: - Parallel and Distributed Constraint Logic Programming Systems,- A Concurrent Constraint-based Unified Framework for Internet Programming, - Parallel Constraint-based Natural Language Processing, and- Parallel Specification and Verification of Real-time Systems,all sharing the common theme of constraint-based technology. For the projects cited, the purchased equipment facilitates: first, the design and implementation of systems that exploit parallelism implicit in logic and constraint programs; second, the distributed network is enabling the prototyping of a unified framework for Internet programming, encompassing symbolic constraints and concurrency; third, parallel constraint software technology will significantly speed up the execution of large natural language processing applications; and fourth, the development of parallel tools for verifying large, complex real-time systems. Parallelism and distributivity are key to accomplishing the goals of these projects. The chosen platform is cost-effective, flexible, and provides a balance between shared memory multiprocessors, which are efficient but not scalable, and distributed networks of processors, which are less efficient but scalable.