The University of Liverpool will use a new Dell high-performance computing cluster to power disease mapping and physics research.
The large-scale Dell HPCC, powered by Intel Pentium 4 processors, will also be used for international public-interest work conducted on the university’s cross-campus Grid network (ULGRID).
The university’s Department of Physics will use the 940-node cluster of Dell PowerEdge 650 servers – expected to be in the top 100 of the TOP500 list of largest supercomputer sites in the world – to simulate the collision of particles to help determine the origins of the universe. ULGRID and the Advanced Institute for Methods and Emergent Systems (AiMeS) will also use the cluster to assist in student and faculty research, to undertake work conducted with the World Health Organization to simulate the spread of epidemics of diseases such as SARS, and to support a planned research Grid with the European Center for Nuclear Research (CERN) in Switzerland.
“The combination of Dell and Intel technology will help to exponentially increase the processing power at our disposal, allowing us to perform larger simulations at a far greater rate and lower cost,” said Themis Bowcock of the Department of Physics. “Our demands were high. We were looking for a cost-effective, reliable cluster solution that would fit within a very restricted space and be deployed as quickly as possible. In addition, we have a limited budget in the department, so we needed a high performance system with low power consumption. Dell’s HPCC solution with Intel Pentium 4 processor met all these requirements.”
The Dell 940-node supercomputing cluster is powered by 3.06 GHz Intel Pentium 4 processors running the Red Hat Linux operating system and is connected by 48 PowerConnect 5224 network switches, providing high-performance switching capabilities. Due to be fully operational in July, the solution will initially co-exist with the university’s original array of 300 PCs and servers, with plans to eventually connect to a wider university science network.
The University of Liverpool is also hosting part of a separate Dell cluster, purchased by a consortium of four universities, including Oxford, Glasgow and University College London. That HPCC provides computing power and data storage for a study conducted by the consortium analyzing particle accelerator data generated and collected by Fermi National Accelarator Laboratory (Fermilab) in the U.S.
Platform Computing Selected For World’s Second Largest Supercomputer at LANL
Los Alamos National Laboratory (LANL) has selected Platform LSF, the foundation of Platform Computing’s Grid solutions, to power its ASCI-Q (Advanced Simulation and Computing Initiative) supercomputer.
LANL’s ASCI-Q supercomputer currently consists of two 1024 4-CPU HP AlphaServer ES45s. The first segment is ranked number 2 on the 2002 Top500 supercomputers listing, with a peak performance of 10 teraflops.
The Q Machine is part of the Department of Energy’s National Nuclear Security Administration Advanced Simulation and Computing (ASCI) program that includes the three DOE defense labs: Sandia National Labs , Lawrence Livermore National Lab and Los Alamos National Lab.
Los Alamos will use Platform LSF to harness the aggregate computing power across the supercomputer, which includes over 8000 CPUs, and to deliver the massive compute power needed to conduct predictive nuclear weapon simulations.
“High-fidelity, three-dimensional computer simulations of unprecedented scale running on ASCI supercomputers such as Q are crucial to the continued certification of the nation’s aging nuclear weapons stockpile in the absence of underground testing,” said Jim Danneskiold of Los Alamos.
“We expect Grid computing infrastructures to become a backbone of national defense and security in the years to come, supporting a seamless, immediate flow of information among various government agencies for optimized security operations,” said Rene Copeland, president of Platform Federal.