About Us

Background

Research Focus

People

Repository

Workshops

Projects

Links

Contact MRI

CCS: Institutes: MRI: Background: Resources of CCS

• Introduction
• Scientific Rationale
• HPC Rationale
• Thrust of CCS-MRI
• Resources of CCS
• Next Steps
• Enabling National Science Goals

Resources of CCS

The goal of Center for Computation Sciences (CCS) is to provide the major source of high performance computing (HPC) resources for computational materials science and condensed matter research communities. These resources are currently some of the best available in the U.S. for materials research. The CCS-MRI provides an opportunity to integrate a critical set of characteristics into a single facility to provide the research community with a capability to develop advanced computational material models. The IBM SP3, with 176 4-processor compute nodes provides the primary platform for CMS simulations and model development. A 64 node Compaq Alphaserver SC with 4 processors per node is also available for code evaluation, optimization, and running smaller applications. The combined peak operating capacity of the CCS is currently 1.5 Teraflops.

A new supercomputer being installed at ORNL's CCS will significantly extend the state-of-the-art resources. The projected performance of 4 Teraflops will place the new machine as one of the world's most powerful civilian supercomputers. The new computer (called "cheetah" after the fastest land mammal) will be housed at ORNL's CCS as part of the DOE's Scientific Discovery Through Advanced Computing (SciDAC) program. Cheetah will be a 24 node IBM Power 4 rated at 4 Teraflops. These nodes have 32 processors each. The machine will have 1 TB of memory and 40 TB of disk. The phased delivery began in December 2001 and will be completed in the first half of 2003.

Electron distribution in the top valence state for (7,0) and (8,0) BN/C nanotubes. Ordered BN/C nanotube structures are predicted to have much better field emission properties than C nanotubes, due to the large pyroelectric field present in the BN segment(s). [V. Meunier, C. Roland, J. Bernholc, and M. Buongiorno Nardelli, Appl. Phys. Lett. 81, 46 (2002).

In addition to these computer resources there are a variety of analysis and development platforms available at CCS. These include a 16 node Compaq AlphaServer, a 32 node SGI Origin, and a 64 node Pentium4. Visualization facilities include immersive technologies such as the ImmersaDesk and CAVE as well as support for advanced presentation graphics and animation. CCS archive and data storage facilities have a capacity of over 300 TB of high speed file storage. The capacity of these data facilities will grow in coordination with the needs of CCS-MRI projects. High-speed networks linking CCS to other national centers and the global Internet provide reliable and efficient services with the ability to move even terascale data sets. CCS is a first tier ESNet node currently operating at OC12.

The CCS-MRI has both staff and student resources to assist in consulting, programming, and performance optimization.