Argonne's Laboratory Computing Resource Center : 2005 annual report.

Bair, R. B. ; Coghlan, S. C ; Kaushik, D. K. ; Riley, K. R. ; Valdes, J. V. ; Pieper, G. P.

Argonne National Laboratory

关键词: Data Base Management;    Computer Networks;    Anl;    Program Management;    99 General And Miscellaneous//Mathematics, Computing, And Information Science;   

DOI  :  10.2172/925354
RP-ID  :  ANL/LCF/RP-59441
RP-ID  :  DE-AC02-06CH11357
RP-ID  :  925354

美国|其它

来源: UNT Digital Library

【 摘 要 】

Argonne National Laboratory founded the Laboratory Computing Resource Center in the spring of 2002 to help meet pressing program needs for computational modeling, simulation, and analysis. The guiding mission is to provide critical computing resources that accelerate the development of high-performance computing expertise, applications, and computations to meet the Laboratory's challenging science and engineering missions. The first goal of the LCRC was to deploy a mid-range supercomputing facility to support the unmet computational needs of the Laboratory. To this end, in September 2002, the Laboratory purchased a 350-node computing cluster from Linux NetworX. This cluster, named 'Jazz', achieved over a teraflop of computing power (10{sup 12} floating-point calculations per second) on standard tests, making it the Laboratory's first terascale computing system and one of the fifty fastest computers in the world at the time. Jazz was made available to early users in November 2002 while the system was undergoing development and configuration. In April 2003, Jazz was officially made available for production operation. Since then, the Jazz user community has grown steadily. By the end of fiscal year 2005, there were 62 active projects on Jazz involving over 320 scientists and engineers. These projects represent a wide cross-section of Laboratory expertise, including work in biosciences, chemistry, climate, computer science, engineering applications, environmental science, geoscience, information science, materials science, mathematics, nanoscience, nuclear engineering, and physics. Most important, many projects have achieved results that would have been unobtainable without such a computing resource. The LCRC continues to improve the computational science and engineering capability and quality at the Laboratory. Specific goals include expansion of the use of Jazz to new disciplines and Laboratory initiatives, teaming with Laboratory infrastructure providers to develop comprehensive scientific data management capabilities, expanding Argonne staff use of national computing facilities, and improving the scientific reach and performance of Argonne's computational applications. Furthermore, recognizing that Jazz is fully subscribed, with considerable unmet demand, the LCRC has begun developing a 'path forward' plan for additional computing resources.