Home / Engineering Computing / New Xtreme-X2 to Advance CFD Research Facility in UK

New Xtreme-X2 to Advance CFD Research Facility in UK

By DE Editors

No sooner had Appro International (Milpitas, CA), provider of high-performance computing (HPC) systems, announced its new Appro Xtreme-X2 Supercomputer, which is based on dual-socket, Quad-Core AMD Opteron processors, than it was awarded a contract capitalizing on its HPC potential.

The contract will culminate in the Xtreme-X2 powering up the HPC resources for a UK-based ING Renault Formula 1 (F1) Team’s Computational Fluid Dynamics (CFD) Centre. So Appro’s recently released Xtreme-X2 Supercomputer will go to work at the Renault F1 CFD Centre — offering a fivefold increase in CFD computing capacity says Appro by adding the ability to run both full-car simulations and aerodynamic testing of critical components (front and rear wings, turning vanes, brake ducts, and fuel tanks, to name a few).

The 38-teraflop (TF) Appro Xtreme-X2 supercomputer that will be delivered to the CFD Centre consists of more than 500 dual-socket, Quad-Core AMD Opteron processor-based compute nodes with more than 4,000 processor cores. As configured, the computer’s peak capability of 38TF is provisioned and managed as a single unified supercomputing system.

"The Xtreme-X Supercomputer addresses the need for higher performance and reliability by offering redundant (Dual-Rail) InfiniBand and Ethernet management fabrics and network switches."
— Jim Ballew, Chief Technology Officer, Appro International, Inc.

The system that Appro is delivering to the UK offers 4.4 terabytes (TB) of high-performance local memory with a usable aggregate bandwidth of more than 8TB per seconds. Each dual processor node has a peak capability of more than 70 gigaflops. The compute nodes in the system are interconnected by a dual-rail double data rate Infiniband interconnect fabric that provides both reduced inter-processor latency and improves system reliability.

"It’s important to note that the HPC market is not only growing, but also maturing, so customers are demanding more of everything — performance, manageability, reliability, availability, flexibility, and TCO,” said Jim Ballew, Chief Technology Officer at Appro International, Inc. “The Xtreme-X Supercomputer addresses the need for higher performance and reliability by offering redundant (Dual-Rail) InfiniBand and Ethernet management fabrics and network switches."

The Xtreme-X2 system, which is scheduled to be shipped and installed by the end of June 2008, comes preconfigured with the Appro Management System, which supports the dual-rail Infiniband network as a load balanced fault-tolerant interconnect and allows the use of shared receive queues with multichannel operation. The software is designed to maximize the bandwidth associated with MPI processes and will support non-stop operation for mission-critical operations, says the release.

The system, which is supported by large memory and visualization nodes equipped with dual graphic processing units (GPU) to allow large panel displays to show simulation results, includes a global parallel file system that is bridged directly to the Infiniband fabric allowing high bandwidth parallel access to the file system from all of the compute nodes.

In general, the Appro Xtreme-X2 features a reliable and scalable architecture with 128 nodes/512 Quad-Core AMD Opteron processors and 4.5TF of computing power in a single 44U equipment rack. It provides 4TB of memory, hot-swappable blades, redundant cooling fans, and efficient “Power-Factor-Corrected” redundant power supplies. Appro’s Xtreme-X2 Supercomputer also offers a new directed airflow-cooling configuration, designed to reduce datacenter floor space by 30 percent while maximizing power and cooling efficiency.

The Appro Xtreme-X2 supercomputer is managed by the Appro Cluster Engine (ACE) management software providing a complete, remote lights-out management solution for the entire system including the interconnect networks, servers, clusters, resource management, and scheduling. ACE, which supports network load balancing and failover, also offers diskless operation with standard Linux distributions and fast boot operations independent of the system size. In sum, Appro says the ACE provides a total management capability for maximum performance and non-stop operation.

In addition, and separately from the supercomputers, Appro will also deliver to the UK Centre multiple XtremeWorkstations based on Quad-Core AMD Opteron processors for the ING Renault F1 Team with an “impressive amount of memory and GPU performance,” states the release.

Appro’s Xtreme-X2, which is based on scalable clusters that offer a cost-effective, highly available, scalable, and energy-efficient architecture, was found to meet the capability computing and high availability requirements needed for the new CFD Centre computational resources.

“Renault is looking forward to the installation of the Appro Xtreme-X2 Supercomputer to enhance our CFD Centre computational resources,” said Bob Bell, Technical Director of ING Renault F1 Team. “Appro not only offered us a cost effective solution but they also improved our required technical specification through better reliability, greater fault tolerance and redundancy as well as more flexibility with regards to system scalability.”

At current conservative development rates, Renault’s expanded facility will provide half of the gains of a fully efficient mature wind tunnel for less than half of the investment. With this added computing capacity provided by Xtreme-X2, the supercomputer will expand the current computational support for Renault F1’s integration into the Renault-Nissan group’s R&D efforts — thus Renault’s compute-intensive CFD Centre will be made available at minimum cost to other business units for research even beyond F1.

Appro points out that this contract marks another milestone for Appro’s supercomputers in international HPC markets, and is in keeping with the company intention to unlock the value of IT through price/performance, balanced architecture, open standards and engineering expertise.

Thus the Appro Xtreme-X2 Supercomputer, which is available now in configurations starting at 64 nodes, with 128 Quad-Core AMD Opteron processors per rack cabinet, and at prices that vary according to system configuration, is being put to good use by Renault. Drawing on the ING Renault F1 Team’s heritage of motorsport innovation that stretches back to Renault’s first Formula 1 race in 1977, the chance of winning the FIA Formula One World Championship with a 100 percent Renault car is that much closer using HPC.

Click here to go to Appro.

Click here to read why DE’s editors chose this as a Pick of the Week.

Click here to go Appro’s webpage about its Xtreme-X2, including wave files.

Click here to see the Specs on the Xtreme-X2.

Click here to access the Features and Benefits on the Xtreme-X2.

Click here to access the Datasheet on the Xtreme-X2.

Click here to view the White Paper on the Xtreme-X2.

Click here to read earlier coverage by DE.

Click here to go tot information on AMD’s Opteron.

About DE Editors

DE's editors contribute news and new product announcements to Desktop Engineering. Press releases can be sent to them via DE-Editors@deskeng.com.