Oak Ridge National Laboratory Boosts Research Capacity with SGI Altix 3000

Oak Ridge National Laboratory (ORNL), part of the U.S. Department of Energy’s research network, has enhanced its high-performance computing environment with the installation of a 256-processor SGI Altix 3000 system. The new platform, equipped with 2 terabytes of globally shared memory, is designed to support next-generation scientific applications that require extremely large datasets and intensive computational processing.

The system was acquired by the laboratory’s Center for Computational Sciences (CCS), where it will be used to accelerate research initiatives spanning environmental modeling, energy systems, homeland security, and advanced scientific discovery. By providing significantly greater processing capacity and memory scalability, the new computing environment enables researchers to simulate complex physical processes and analyze massive data volumes that were previously difficult to handle efficiently.

Built with 256 Intel Itanium 2 processors, the Altix 3000 delivers approximately 1.5 teraflops of computational performance. Its architecture supports shared-memory processing at large scale, allowing scientists to run highly demanding applications such as climate modeling, genetic and biological simulations, materials science research, and fusion energy studies. The shared-memory design is particularly valuable for applications that require rapid communication between processors or simultaneous access to extremely large datasets.

To complement the new computational platform, ORNL has also implemented more than 12 terabytes of high-performance RAID storage using SGI’s Total Performance 9100 system. The Fibre Channel–based storage infrastructure supports both direct-attached and storage-area-network configurations, ensuring that the large data streams generated by scientific simulations can be stored, accessed, and processed efficiently. The system operates within an open, Linux-based environment, aligning with the laboratory’s strategy of deploying scalable, standards-based computing platforms for high-performance applications.

The Center for Computational Sciences, established in 1991, plays a central role in evaluating emerging computing architectures and collaborating with universities, government laboratories, and industry partners to develop scalable scientific software capable of leveraging advanced computing systems. Through partnerships with institutions such as Georgia Tech, Virginia Tech, the University of Tennessee, Duke University, Florida State University, and other academic collaborators, ORNL plans to assess the capabilities of the Altix architecture for large-scale engineering and scientific modeling projects.

The Altix 3000 platform reflects the broader evolution of high-performance computing systems that combine advanced processor technologies with open operating environments. The integration of Intel Itanium processors, Linux operating systems, and scalable cluster technologies has enabled dozens of scientific, engineering, and industrial software applications to be optimized for 64-bit Linux environments, supporting increasingly complex research workloads across multiple sectors.

Oak Ridge National Laboratory, managed by UT-Battelle for the U.S. Department of Energy, conducts multidisciplinary research focused on expanding scientific knowledge, advancing clean energy technologies, improving environmental sustainability, and supporting national security initiatives. The deployment of the new Altix 3000 system further strengthens the laboratory’s capacity to conduct large-scale computational research and maintain leadership in advanced scientific computing.

Silicon Graphics, Inc. (SGI), headquartered in Mountain View, California, is a global provider of high-performance computing, visualization, and data storage technologies designed to enable major breakthroughs in science, engineering, and digital media. The company’s advanced computing systems continue to support research organizations, energy companies, and scientific institutions seeking to address complex computational challenges through scalable supercomputing platforms.