Appro today announced the completion of the first of four super compute Linux clusters for Lawrence Livermore National Laboratory. Named "Rhea," the Infiniband(TM) interconnected cluster boasts 576 Next-Generation AMD Opteron(TM) 8000 Series processors with peak processing power of 22 teraflop/s (trillion floating operations per second). The Rhea server was brought online on October 23, 2006 and just completed its testing phase.
The three remaining clusters named Zeus, Atlas and Minos are scheduled for completion by the end of Q1 2007. Each will feature the Appro Quad XtremeServer solution used to create Rhea. When the additional supercomputing clusters are complete, they will offer LLNL scientists 2,592 4-socket / 8-core nodes and approximately 100 teraflops/s of processing capacity, ranking the group of clusters in the top three supercomputing resources at LLNL.
According to LLNL, the new dedicated compute-intensive and results- oriented clusters will provide their scientists the power, throughput and agility they require to handle sophisticated multi-programmatic, native 64-bit applications. Typical uses for LLNL supercomputer clusters are processing applications that support nuclear stockpile stewardship, climate modeling, protein folding, material modeling, dislocation dynamics, atmospheric and ground flow and earthquake simulations. Before these new Linux based supercomputer clusters were designed, applications of this nature required extensive job queuing and lengthy scheduling. Now, many of these jobs can be processed a lot faster.
"We need supercomputing resources that power scientific discovery," said Mark Seager, who leads LLNL's program to develop new platforms. "These supercomputing resources were created to meet the exacting requirements of the National Nuclear Security Administration (NNSA) responsible for the safety, security and reliability of the nation's aging nuclear deterrent. The U.S. Department of Energy and NNSA are committed to the acquisition of cost effective systems that can be rapidly deployed to benefit a broad range of scientific research in the national interest."
"Participating in the creation of the next generation of supercomputers at LLNL is right where Appro belongs," said Daniel Kim, CEO of Appro. "Through the creation of these HPC clusters, we have proven once again that performance; integration and flexibility are not only the requirements of HPC clusters but also of those who create such environments. Appro's extensive HPC experience and flexibility in delivering customized configurations, combined with customer focus and a strong technology alliance with AMD met LLNL's unique requirements."
"LLNL's Rhea cluster is further proof that 4-socket servers are optimized for more than just databases today," said Kevin Knox, vice president, Worldwide Commercial Business, AMD. "AMD and partners like Appro have helped bring volume economics to the 4-way server market. Customers like LLNL are recognizing they can get the price/performance-per-watt and scalability, particularly when they see the seamless upgrade path to native quad-core processing that only AMD64 with Direct Connect Architecture can deliver."
To learn more about these supercomputing clusters and other industry leading products, visit Appro's booth 1634 at SuperComputing show from 11/13 to 11/16 in Tampa, Florida.