Our Company

Headache-free HPC!

Advanced Clustering Technologies is a leading provider of high performance Beowulf clusters, Linux-based servers and workstations.
Advanced Clustering — YOUR source for turnkey high performance computing hardware, software, and services.

Customer testimonial

"I have found Advanced Clustering to be a good source for integrated Linux systems. Service has been great and extremely knowledgeable... We can call and get immediate answers to our questions."  --Phil N. at NASA Goddard Space Flight Center

Find out even more reasons to purchase from us

 
Home Newsletters June 2012

June 2012

 

alt

 

 

   June 2012       

Contact

 

Phone

Toll-free: 866.802.8222

International: 913.643.0300
 
Email
 
 
 
 
 
 

 

alt

 

We want to hear your story!

 

If you're a user of an Advanced Clustering Technologies cluster, we would love to find out what you're doing with it!

 

When we sell a cluster, we have a general understanding of how it will be used, but we don't always get to hear the details about the research projects our customers are conducting or what sort of discoveries are made. We'll admit we're curious!

 

Please consider dropping us a quick note to let us know all about your cluster experience at  This e-mail address is being protected from spambots. You need JavaScript enabled to view it . We may publish it in an upcoming newsletter, and if we do, you will receive FREE SHIPPING on your next order with us, no matter how large or small!

 

 

Intel Xeon "SandyBridge" Processors Now

Part of TOP500; U.S. Back in First Place

 

Released just this past March, the Intel Xeon E5 "SandyBridge" processors have found their way into the most powerful supercomputers in the world.

 

The June 2012 edition of the TOP500 supercomputer ranking list now includes clusters based on the E5 chips, making these new processors the fastest-adopted technology in high-performance computing. These new processors can be found in 45 of the TOP500 systems, three of them offering petascale-class computational power.

 

These Xeon E5-based supercomputers are located in the United States, Europe, India, China and Australia, where they are assisting researchers in fields such as computational chemistry, fluid dynamics, astrophysics and medicine.

 

For information on Advanced Clustering's own Intel SandyBridge E5-series offerings, please click here: www.advancedclustering.com/products/intel-xeon-e5-systems.html

 

In other TOP500 news, the United States is now back at number one on the list with the U.S. Department of Energy's "Sequoia," a nuclear-explosion simulator built to eliminate the need for real-life nuclear testing. Housed at the Lawrence Livermore National Laboratory in California, this IBM BlueGene/Q system uses 1.57 million cores of Power BQC processors that are capable of processing 16.32 petaflops per second.

 

(To put this in perspective, it would take 320 years for every single person on Earth to perform the same number of calculations Sequoia can process in one hour.)

 

The U.S. last held the top spot on the list in 2009 with the Dept. of Energy's "Jaguar" supercomputer, but lost this position in 2010 and 2011 to systems based in China and Japan.

 

To see the full list, please visit www.top500.org/lists/2012/06.

 

The upcoming Intel Xeon Phi: 

Supercomputer in a chip

 

Intel has announced the development of a new "coprocessor" chip that it believes will dramatically increase the parallel processing power of clusters. The Xeon Phi is a many-integrated-core chip, or "MIC," that will work in tandem with traditional processors, similar in concept to a GPU.

 

Intel has dubbed the Xeon Phi a "supercomputer in a chip," because this coprocessor offers one teraflop of performance, the same as the first Intel-based supercomputer built in 1997. The difference is that while the 1997 system featured more than 9,000 processors in 72 server cabinets, the Xeon Phi will occupy just one PCIe slot in one server.

 

This new coprocessor is not just theoretical: Intel has built a supercomputer with a beta version of the Xeon Phi, and this system was listed as number 150 on the latest TOP500 list. Wide distribution of this new chip is scheduled for sometime in 2013.

 

You can read more about the Xeon Phi here: blogs.intel.com/technology/2012/06/intel-xeon-phi-coprocessors-accelerate-discovery-and-innovation

 

For upcoming news on Xeon Phi offerings from us at Advanced Clustering, keep reading this newsletter and visit our webpage at www.advancedclustering.com.

 

 

Shipping soon: NVIDIA's Tesla Kepler K10 GPU

 

NVIDIA recently announced that its new Kepler K10 GPU, part of the Tesla GPU family, has been released and will be available for shipping in the next few weeks.

 

The new K10 GPU is the first of two Kepler GPU products that promise greater performance and energy efficiency than their previous Tesla counterparts, such as the M2090.

 

The K10 is designed primarily for single-precision floating-point calculations, so its current uses are somewhat limited, although users who require fast single-precision computation will appreciate the jump from 1.3 to 4.58 teraflops per second.

 

The second in the Kepler line, the K20, is scheduled to be released at the end of this year. This GPU will offer three times the double-precision processing speed of the M2090, plus the ability to dynamically generate new threads and allow multiple CPU cores to simultaneously access the CUDA cores on one GPU, thus making it ideal for parallel computing applications.

 

Want to know more about these new GPUs or GPU utilization in general? If so, contact us at  This e-mail address is being protected from spambots. You need JavaScript enabled to view it , and we'll be happy to answer your questions.

 

GPN Annual Meeting Wrap-Up

 

Advanced Clustering Technologies proudly attended the 2012 Great Plains Network Annual Meeting in Kansas City, May 30-June 1. This conference brings together network and cyberinfrastructure leaders from the region's top research and education universities to explore the critical technology challenges facing the industry.

 

For the fourth consecutive year, supercomputing played an important role in this conference, which gave us the opportunity to discuss the current state of HPC with many of the conference attendees, including representatives of several area supercomputing centers (Kansas State, University of Missouri, University of Arkansas, Oklahoma State and University of Oklahoma).

 

As the only HPC computing company on hand, we were in a unique position to provide in-depth information on the latest CPU architectures and network interconnects and offer insight into future HPC trends.

 

This was our second time attending this meeting, and we always welcome the opportunity to renew acquaintances with such influential members of this community.

 

For more information on the Great Plains Network, please visit

www.greatplains.net/display/Home/Great+Plains+Network.