MRI: Acquisition of a Computational Science and Engineering Parallel Cluster

提案编号：CNS 09- 23050 PI：Jones，Jim; Jacher，Steven M.;佛罗里达理工学院 墨尔本，佛罗里达州32901- 6975标题： MRI/采集：获取计算科学与工程并行计算机项目建议：该项目，收购一个新的计算科学与工程集群，旨在为教师和学生提供定期访问现代化的校园并行计算机进行研究。该项目服务于机械和航空航天工程，物理和空间科学，海洋和环境工程，海洋学和数学科学，旨在继续进行涉及各种学科并行计算的研究。由于其年龄，2001年收购的现有Beowulf集群已被证明不足以进行当前的研究，并寻求更强大的计算资源。该仪器由48个节点组成，每个节点有2个四核2 GHz CPU。这台机器的计算能力将使海洋建模、空间天气和材料科学等应用领域所需的大规模模拟成为可能。在这些领域，研究人员有模拟代码，但需要增加计算能力来进行所需的运行。这台新机器预计将允许对重要应用进行更逼真的模拟，例如空间辐射对宇航员和卫星上电子元件的影响。其他应用程序需要增加的计算能力，并开始并行化他们的代码。更广泛的影响：这个项目有助于增加校园访问的学生发展他们的并行编程技能。基于过去的经验，高性能和并行计算的职业生涯将受到鼓励。学生将在该领域接受更深入的培训，并行计算将继续成为该机构研究的一部分。

Proposal #: CNS 09-23050PI(s): Jones, Jim; Jacher, Steven M.; Wang, Ke-Gang; Zhang, Ming Institution: Florida Institute of Technology Melbourne, FL 32901-6975Title: MRI/Acq.: Acq. of a Computational Science and Engineering Parallel ClusterProject Proposed:This project, acquiring a new computational science and engineering cluster, seeks to provide the faculty and students regular access to a modern, on-campus parallel computer for their research. Servicing Mechanical and Aerospace Engineering, Physics and Space Science, Marine and Environmental Engineering, Oceanography, and Mathematical Sciences, the project aims to continue conducting research involving parallel computing in a wide variety of disciplines. Due to its age, an existing Beowulf Cluster acquired in 2001 has proved insufficient to conduct the current research and more powerful computing resources are sought. The proposed instrument consists of a 48 node cluster where each node has 2 quad-core 2 GHz CPUs. The computing power of this machine will enable the large-scale simulations needed in the application areas of ocean modeling, space weather, and material science. In these areas researchers have simulation codes, but need increased computing power to do the required runs. The new machine is expected to allow greater fidelity simulation on important application, such as the impact of space radiations on astronauts and electronic components on satellites. Other applications need the increased computing power and are beginning to parallelize their codes.Broader Impacts: This project contributes to increase on-campus access for students to develop their parallel programming skills. Building on past experience, careers in high performance and parallel computing will be encouraged. The students will be trained more in depth in the area and Parallel Computing will continue to be part of research at the institution.