INDUSTRY

AUSTIN, TX -- The Texas Advanced Computing Center (TACC) at the University of Texas at Austin (UT) will be receiving a donation of high-end hardware from Intel Corporation that will enhance the computational research and training at the Center as well as providing new resources for grid development in the state of Texas. The emerging grid infrastructure will ultimately benefit scientists nationwide by allowing a fast, efficient way for them to process and share data. "TACC is extremely grateful for this generous donation from Intel. The TACC staff is eager to begin evaluating, developing, porting, and installing network, security, and grid software technologies that use this hardware to enhance the usefulness of TACC resources to users at UT and across the country," said Jay Boisseau, director of TACC. "The grid activities we conduct for the University and State of Texas will include much software from the National Partnership for Advanced Computational Infrastructure [NPACI] and therefore will help build expertise that assists NPACI in its construction of the TeraGrid." Grid technologies tend to be "behind the scenes" from a computational science perspective while greatly enhancing the capabilities of researchers by integrating separate resources more tightly and enabling the use of multiple systems more effectively. TACC grid computing research and development will provide an opportunity for UT students to participate in this fast-growing technology area. "Intel is especially pleased with the recent equipment contribution to TACC because this equipment will support graduate students in their research on primary code development, parallel computing, and performance optimization," said Kimberly Sills, Academic Relations Account Manager at Intel. "In addition, TACC, as an NPACI resource partner, plays a key role in the success of NPACI and its support of the TeraGrid, which Intel is helping to build," said Sills. The new hardware donated to TACC includes: -- Two workstations: Intel Xeon 1.7 GHz, Intel 860 Chipset/400MHz system bus, 2 9-pin serial connectors, 16550-compatible, 25-pin parallel connector (bi-directional), 1-4XAGP Pro110 slot, 3-32 bit/33 MHz & 2-64 bit/66 MHz PCI slots, 2-IEEE 1394 ports -- One dual-processor server: Intel Pentium III Xeon 1GHz, 133MHz front side bus, 7 PCI slots (2 x 64-bit/66MHz (hot-plug), 2 x 64-bit/33MHz (hot-plug), 2 x 64-bit/33MHz, 1 x 32bit/33MHz -- Four single-processor servers: Intel Pentium III Xeon 1GHz, 133MHz front side bus, 7 PCI slots (2 x 64-bit/66MHz (hot-plug), 2 x 64-bit/33MHz (hot-plug), 2 x 64-bit/33MHz, 1 x 32bit/33MHz) -- Three Gigabit Ethernet NICs: IntelR PRO/1000 F Server Adapter -- Three Gigabit Ethernet Switches: NetStructure 470F GB Switch TACC plans to use the new network equipment to link TACC systems at higher bandwidths and the computing hardware to provide network, UT, and grid software services. The network hardware will be used primarily to enhance the production capabilities of the TACC infrastructure and the developing TACC grid; the computing systems will be used for hosting both research and development software and for production software that results from those research and development efforts. For example, the most powerful computing systems on the UT campus are in the TACC main facility, while the immersive scientific visualization laboratory is nine miles away. The largest data storage and archival facility is in the TACC machine room, so a process of simulation-visualization-storage requires multiple data transfers of potentially enormous files. As UT researchers develop their own local clusters for code development and small scale simulation, there is often need to move data from these systems to the Vislab for analysis and the TACC data archive. TACC will use NPACI-developed software, including the SDSC Storage Resource Broker, which provides a "data grid," a single logical name space extending over a heterogeneous Grid of computing and storage systems, and the Network Weather Service, a collection of network performance data for use in making Grid scheduling decisions, such as scheduling of data transfers for simulations. For additional information visit www.tacc.utexas.edu/ , www.npaci.edu or www.intel.com