ASC Supercomputers
Some of the world’s fastest supercomputers are among ASC’s accomplishments in advanced computing. However, it is not all about speed. Each new system is engineered to bring certain capabilities to bear on the problems of modeling and simulation that will enhance the overall goals of the Science-Based Stockpile Stewardship Program. In doing so, ASC will continue to challenge the state of art for computer scientists and manufacturers with systems such as those described below:
BlueGene/L
The unique BlueGene/L “compute card” consists of two nodes, each containing two central processing units (CPUs) including four megabytes of dedicated memory. Also mounted on the circuit board are a memory controller module and nine separate memory chips, similar to those in standard home and office desktop computers.
Red Storm
The Red Storm Architecture ensures a balance in all aspects of the architecture: a partitioning model that provides a level of optimization for each of the functional features required of a large-scale system, an attention to building the system as a single integrated machine and an applied engineering discipline to ensure a high degree of reliability for application codes running across the entire machine. This integrated approach provides for an extreme level of performance and reliability that can not be obtained in systems based on commodity components and disparate integration technologies not necessarily designed for the rigors of supercomputing.
Roadrunner
Los Alamos National Laboratory has completed the first two of three phases of the Roadrunner Project. Phase 3 of the project will install the complete Roadrunner system, which is a cluster of standard Linux systems using AMD dual-core Opteron™ processors accelerated with new enhanced double precision IBM PowerXCell 8i™ chips. The Roadrunner Phase 3 system is expected to reach a sustained 1 petaFLOPS in the summer of 2008 making it more than twice as fast as the current fastest supercomputer. Los Alamos scientists are leading the way in hybrid computing and are the recognized experts in this area.