By Anthony J. Lockwood
Dear Desktop Engineering Reader:
Supercomputing has and will continue to revolutionize the playing field for practitioners of high-level analysis in industries as varied as automotive, aerospace, and oil and gas exploration. Until recently, however, supercomputing was restricted to outfits that could foot the hardware and administrative costs. But over the past couple of years, the cost of getting into supercomputing has come down to the point where small- and mid-sized organizations can get onto the field. One of the leaders in this price/performance breakthrough has been Appro. And they’ve just made the entry-point more attractive as part of an expansion of their Xtreme-X Supercomputer line and cluster management software.
The Appro Xtreme-X Supercomputer offers a scalable architecture that groups high-performance servers together into a unified, fully integrated Scalable Unit (SU) that can be provisioned and managed as a stand-alone supercomputer. The key point there when talking entry-level supercomputing—high-end for that matter—is that you can join servers together in a building block fashion easily as you go along. So, the Xtreme-X Supercomputer provides a growth path for you.
Let me step back here and give you some basics. The Xtreme-X Supercomputer is based on the multi-core (six cores if you’re an incorrigible counter) Intel Xeon Processor 5600 Series. It’s available with interconnect fabric options for both single-rail InfiniBand Network or dual-rail configurations. Standard features include Appro’s well-regarded power and cooling efficiency, memory capacity of up to 6TB per rack, and a per rack peak performance of 9TF.
But an equally important key to supercomputing computing is harnessing its power. That’s where Appro’s ACE (Appro Cluster Engine) Management software comes in. ACE provides Xtreme-X Supercomputers plug-and-play functionality that let’s you build out your system by adding more server hardware that ACE discovers and places online automatically. Still, ACE does more than that. It offers diskless booting of compute servers and the ability to independently configure and provision individual clusters to enable fine-tuned workload and energy management. ACE enables applications to locate an optimized cluster environment, which means you can go about managing the global compute environment and energy consumption without fussing over where to put jobs.
Now, the Xtreme-X Supercomputer EL Series is engineered as an entry-level system. It has optional NVIDIA Tesla graphics support. It’s joined by the MR Series for mid-range performance and the HE for high-end, high-availability performance. You can read about the new Xtreme-X Supercomputer configurations and the enhancements to ACE in today’s Pick of the Week. Then, hit the link to see the videos. If you can watch only one, then watch the first video on ACE. (Note: I had trouble getting Firefox to play this video, but Explorer fired it right up.) The thing to know about Appro is that it has your supercomputing requirements covered now and as you grow. Learn about them.
Thanks, Pal. — Lockwood
Anthony J. Lockwood
Editor at Large, Desktop Engineering