Administrative and Computational Core Facility

点击翻译按钮获取中文摘要

A.ADMINISTRATION. The Program Project Grant Director, Dr. Segrest, is Director of the Atherosclerosis Research Unit (ARU), the administrative home for the program project. He also is Director of the Center for Computational and Structural Biology at UAB. He will be responsible for planning of the overall direction of the program and for quality control. One of the major roles of the Director will be to foster interactions and to stimulate new ideas and directions within and between the various projects and core units. Systems management support and programming support for the Program Project will be provided by Martin Jones, MS. Mr. Jones manages the computing section of the ARU and assists investigators with programming solutions. He also provides assistance in troubleshooting, pricing, and computer replacement/upgrading. He is also currently working on two research problems: i) He is updating the "Wheel Programs" suite (consisting of the programs Locate, Wheel, Helnet, and Combo) to use a more portable and usable interface. He is also working to make these programs more readily available. (See http://www.aru.uab.edu/ resources.html) ii) He is working with Dr. Gilbert Weinstein of the Department of Mathematics at UAB to computationally model the dynamic physical and differential geometric evolution of HDL-like disc particles. Most of the clerical and desktop publishing services, correspondence, etc. will be provided by LaDonna Bangham. She will prepare and submit purchase orders, track expenditures, and make the travel arrangements for the program faculty. She provides support for faculty and staff involved in all aspects of the PPG. Ms. Bangham will be responsible for coordination of the travel arrangements for the PPG External advisory Committee and external speakers for the ARU Scientific Seminars. An ARU Internal Advisory Committee, independent of the program project investigators, provides periodic scientific and administrative advice to the Director of the program project/ARU. The PPG Internal Advisory Committee will meet monthly to provide information and advice to the Director. The PPG External Advisory Committee will consist of four extramural scientists to be appointed to overlapping three year terms who will perform a site visit yearly to review individual components of the program project for the Director. B.COMPUTATION Shared high performance computing facility used by this PPG. The primary computational resource for Project 1 of this PPG is a single rack UAB Blue Gene. The UAB Blue Gene, an IBM Blue Gene/L, is a uniquely designed massively parallel supercomputer. A single Blue Gene/ L rack has 1024 nodes, each node having two processors and 512MB of main memory. The 2048 processors in the Blue Gene/L rack are tightly integrated in one form factor using five proprietary high-speed interconnection networks, producing a peak performance of 5.6 Teraflops. There are 13 Terabytes of disk storage available to this cluster. The Segrest group is currently scheduled for 50% minimum usage of the UAB Blue Gene, with on-demand utilization of 100% of the 2048 processors of the UAB Blue Gene. Secondary Computing resources are provided by three commodity clusters. Coosa: A DELL Xeon 64-bit Linux HPC Cluster that consists of a 128-node DELL PE1425 computer that has dual Xeon 3.6GHz processors with either 2GB or 6GB of memory per node. There are 4 Terabytes of disk storage available to this cluster. This cluster is rated at more than 1 Teraflops computing capacity Cheaha: A Verari Opteron 64-bit Linux HPC Cluster that consists of a 64-node dual AMD Opteron 242 computer with 2GB of memory each node. Both Coosa and Cheaha clusters use Gigabit Ethernet inter-node network connections. Olympus: A 128-node (256- processor), 1.4Teraflop-rated grid node that is similar to Coosa The Coosa and Cheaha clusters, as well as the DAB Blue Gene, are supported by Dr. Bharat Soni's group in the School of Engineering. Olympus support is provided Dr. Tony Skjellum's department - Computer and Information Sciences. The Segrest research group is flexibly scheduled for 264 (64-bit) cpu from these clusters at any one time.