14 DAY TRIAL // Appro has announced recently that the company has successfully deployed the ApproXtreme-X supercomputer, also known as Mustang, based on AMD's Opteron 6100-series processors at the Los Alamos National Laboratory.
According to Appro, this deployment represents approximately $10 Million in production computing capability and will be used for a variety of applications, including ocean, wildfire, plasma physics, materials and nuclear energy research.
The ApproXtreme-X supercomputer includes two 12-core AMD Opteron 6100 series processors per node in order to deliver a total of 38,400 computing cores, capable of providing up to 353 Tflops of performance.
The system is designed to optimize floor space, power, and cooling requirements as it utilizes an air-cooling system paired together with platinum-rated power supplies and high-efficient fans, all integrated with the supercomputer using a console management interface.
All the nodes installed in the machine are interconnected through a single-rail 4x Mellanox QDR IB (Quad Data Rate InfiniBand) link and integrated with the Tri-Laboratory Common Computing Environment (CCE) software stack built from Red Hat Enterprise Linux distribution and additional 3rd party and open source software.
“Scientific computing is a core capability of Los Alamos National Laboratory (LANL) and the addition of the ApproXtreme-X Supercomputer based on the AMD Opteron processor will provide a broad array of science-based prediction modeling " said Andy White, Deputy Associate Director for Theory, Simulation and Computation.
“By combining observations, data, and theory with complex modeling, extremely fast computers and visualization capabilities we can achieve a better understanding of the physical world on all scales; from the cosmically large to the infinitesimally small.”
Appro hasn't commented on the upgrade capabilities of the Mustang supercomputer, but there is a strong possibility that the system can take AMD's upcoming Opteron 6200-series processors which can pack up to 16 computing cores.