Scientific Computing Research Environments for the Mathematical Sciences

NSF提案DMS-0215491 PI：Fairweet、Bialecki、Heremann、Rockwood和Wang。摘要科罗拉多矿业学院数学和计算机科学系的新研究计算设施将通过为教师和研究生提供最先进的高性能计算环境来增强其研究计算基础设施。它将包括14个双处理器工作站，每个工作站具有1 GB内存和40 GB硬盘驱动器，以及一个具有2 GB内存和108 GB硬盘驱动器的双处理器服务器。这些机器将配置最新版本的RedHat Linux。这些机器将与MCS现有的一些机器相连，以扩展用于高性能计算的本地分布式网络。代码开发将在Linux/Unix环境中完成，编程主要使用C/C++、Fortran 90/95和Java。此外，MCS还拥有数学、Maple、sPlus和MatLab的站点许可证，以提供全套符号计算能力。需要新的研究计算环境来解决MCS研究人员正在考虑的复杂问题，特别是在矩阵分解算法、偏微分方程组的节点配置、非线性偏微分方程组和格子的符号软件开发、有限元数据插值法和用于建模的广义最小二乘插值法等领域。其中九台机器将构成计算机实验室的核心，供毕业生使用，帮助教职员工进行研究。在所提出的系统下，这些机器可以用于研究人员的个人计算，也可以组成一个分布式计算网络。购买双处理器计算机而不是单处理器计算机将使分布式计算网络中的处理器数量增加一倍，而成本增加不到20%。分布式网络将用于代码开发、编译、调试和生产运行。

NSF Proposal DMS-0215491PIs: Fairweather, Bialecki, Heremann, Rockwood, and Wang.ABSTRACTThe new research computing facility in the Department ofMathematical and Computer Sciences (MCS) at the ColoradoSchool of Mines will enhance its research computinginfrastructure by providing faculty and graduate studentswith a state-of-the-art high performance computing environment.It will comprise 14 dual processor workstations, each with1 GB memory and 40 GB hard drive, and a dual processor serverwith 2GB memory and 108GB hard drive. The machines will beconfigured with the latest version of RedHat Linux. Thesemachines will be linked with some currently existingmachines at MCS to extend the local distributed network forhigh performance computing purposes. Code development willbe completed in the Linux/Unix environment with programmingprimarily in C/C++, FORTRAN90/95, and Java. In addition, MCShas site licenses for Mathematica, Maple, Splus, and Matlabto provide a full suite of symbolic computing capabilities.The new research computing environment is required to solvethe complex problems being considered by researchers in MCS,specifically in the areas of matrix decomposition algorithms,nodal collocation for partial differential equations,development of symbolic software for nonlinear partialdifferential equations and lattices, data interpolation byfinite element methods, and generalized least squaresinterpolation for modeling. Nine of the machines will formthe nucleus of a computer laboratory to be used by graduatestudents assisting faculty members in their research. Underthe proposed system, the machines can be used for individualcomputing by researchers, and also comprise a distributedcomputing network. The purchase of dual processor computersrather than single processor computers will double the numberof processors in the distributed computing network at a costincrease of less than 20%. The distributed network will beused for code development, compilation, debugging, andproduction runs.