Intel has delivered on "Moore's Law" using dual-core processing to build a 128-node High-Performance Computing (HPC) cluster that delivers theoretical peak performance of 3.2 teraflops and sustained performance of over 2.1 teraflops. Based on off-the-shelf technologies, including the next-generation dual-core Intel Xeon processor and an InfiniBand interconnect, the cluster represents a new era that rapidly increases performance while reducing or holding steady the requirements for power, heat and floor space. Industry collaborators and end-users can access the machine through the Intel Remote Access Service and use it to test-drive their codes and accelerate their move to Intel multi-core computing.

It takes a comprehensive strategy to scale high performance computing (HPC) capabilities, while simultaneously containing power and cooling costs. New Dual-Core Intel Xeon and Intel Itanium processor-based servers offer a critical new resource, delivering dramatic increases in performance, price/performance and energy-efficiency across a broad range of HPC applications. Read about this and other Intel advances that can help you increase density, reduce costs and scale capacity in your existing facilities.

This paper explores the market adoption of and customer value proposition and adoption plans for the Quad-Core Intel Xeon processor 5300 series since its market introduction in 2006, focusing on workloads that are designed to run on multicore and multisocket platforms. In addition to discussing the range of workloads, IDC has interviewed customers who are adopting the Quad-Core Intel Xeon processor 5300 series to learn about the IT requirements that they are addressing and the deployment patterns found in those sites.

This document presents a comparative analysis of the performance of LS-DYNA, a finite element analysis application, using two interconnect and fabric technologies: Gigabit Ethernet and Cisco InfiniBand.

The InfiniBand standard supports single, double, and quadruple data rate that enables an InfiniBand link to transmit more data. This paper discusses the characteristics of single data rate (SDR), double data rate (DDR), and quad data rate (QDR) InfiniBand transmission with respect to performance, transmission distances, and cabling considerations.

Traditionally, parallel applications have run on monolithic supercomputers that have been prohibitively expensive for many companies to acquire and operate. A recent development that uses much the same principles as traditional supercomputers are HPC clusters. HPC clusters are made up of multiple, sometimes many thousands, of industry standard computers that use cluster software and high-performance network interconnects to run parallel applications at a fraction of the cost of traditional supercomputers.

Over the last 40 years, the use of Computer Aided Engineering (CAE) has increased by several orders of magnitude in industry and research laboratories, largely due to the impressive advances in computing architectures as well as in the algorithmic techniques created to exploit these architectures. Each major innovation in the computing industry has directly enabled CAE practitioners to solve more realistic and complex engineering design and simulation problems, resulting in better products faster. This investment by the community in adapting CAE applications to take advantage of newer computing architectures has paid off handsomely.

The organization faced several major challenges in providing IT resources for high-throughput biomedical research, including providing high-performance computing capabilities that allow TGen's scientists to analyze large volumes of complex data very quickly.

Achieve highest possible computational performance with limited budget and floor space; resolve issues related to concentration of heat output from ultra-dense blade servers.

This week I'd like to take a look at the latest survey results. While the participation was not all that I wanted, it is enough to make sufficiently vague statements about the HPC cluster market.

What can we learn from the ants and bees? Perhaps something we can take something from the ants that may be helpful for clustering. If nodes were cheap and plentiful, then who cares? It would be kind of like stepping on ants: there always seem to be more.

Trying to get a handle on the HPC community and market has always been difficult. The professional market forecasters seem to paint a rosy picture for HPC over the next five years. But, when I talk to people in the market, I often get a different take.

Doug Eadline takes a break from his ranting about multi-core CPUs to rant about another technology that gives him fits: Virtualization.

It's time for the yearly batch of retrospectives and predictions. Count me in! Let's see, the big thing of 2007? Well, that had to be multi-core. And, the big prediction for 2008? Why that would be multi-core, once again. There, I'm done. Enjoy your year.

As 2007 fades away, I thought I would reflect on some of the HPC events of the last twelve months. Having thought about it, though, nothing really stands out in my mind as a big breakthrough or new paradigm shifting technology.

Setting up a cluster can be trying enough, and maintaining it can be even more difficult. The sheer number of nodes involved in a large cluster can be daunting, as can users' expectations for quality of service. To make life easier, Troy Baer provides a tour of tools that every cluster admin should know about.

With all due respect to Forest Gump, defining fast is becoming a bit harder these days. And, yes, it has to do with multicore.

Suffering from multi-core anxiety syndrome (MCAS)? Need honest answers to you infrastructure questions? Want to pick the brain of an expert? Join the live discussion and let our panel put you on the right track.

Learn how graphics processing units are pushing the HPC envelope.

Earlier this year, my wife and I decided to finish our basement. Great! I thought I could finally set up a real office. I had just one thing to do before my office would become a reality — deal with the ghosts of systems past.

After a fun but exhausting week at SC07, HPC editor Doug Eadline returns to report on the show.

In the last installment of our Parallel Platters series, Jeff Layton looks at the next generation of parallel file systems: Object Based File Systems.

I am standing in the middle of SC07 (Supercomputing 2007). SC07 is THE HPC event of the year. 318 exhibitors have made the trek to Reno, Nevada for SC07, and scores of attendees are here for the week-long conference. If you're attending the show, you're probably not reading this because you're either completely exhausted, back-logged, depressed from losing your money at the casinos, still trying to get your shampoo back from the TSA agent at the airport, or some combination of the above.

Heading to SC07 (Supercomputing 2007)? To get prepped for the show, listen to HPC editor Doug Eadline and editor-in-chief Joe Brockmeier discuss what to expect from SC07, what to see while you're there, and where to get free beer.

I find Pablo Picasso's famous response on computers, "Computers are useless. They can only give you answers," interesting and provocative. I also believe there is a grain of truth in that statement — namely, that it's really important to ask the right questions of computers. I have a hard time with the useless part, however.

When comparing CPUs, you have to compare apples to apples. Doug Eadline compares the performance of AMD Opteron and Intel multicore processors to see which CPUs provide the best performance per core.

Continuing the discussion of cluster-ready filesystems with an overview of traditional parallel file systems, which allow clusters to contact multiple storage devices directly rather than communicating with storage through a gateway. All you ever wanted to know about parallel file systems, and then some.

Back in the good old days of single core processors, when HPC clustering was in its infancy, getting the application as close as possible to the hardware was very important. In many cases, it still is. Communication between nodes could take place through the operating system by using TCP/IP or outside the OS using a userspace zero-copy protocol. With the exception of pinning down memory, the userspace protocol totally removes the OS from the communication. The result of userspace communication is better application performance thanks to better latency and throughput.

A story floating around the Internet recently claims that the Storm email worm has created the world's largest supercomputer. I believe the genesis of this claim was this post on a security Web site. Are you amazed? Do we need James Bond to ferret out the evil villains? Hardly. Perhaps Austin Powers can handle this one.

What Drives Performance in HPC? That's a good question. What does drive performance in HPC? On a qualitative basis its easy to answer. A faster processor and memory. More memory. A better network or disk I/O subsystem. Unfortunately, those answers are rarely specific enough when faced with purchase decisions for a Linux cluster.

The free sharing of ideas, successes, and failures around Linux clustering was, in part, responsible for the rapid growth of Linux clusters. But, where are all the HPC user groups now?

High Performance Computing (HPC) clusters are easier, and cheaper, than ever to put together. If you have an interest in pulling together your own cluster, or maybe you just want to understand more about cluster technology, it’s necessary to grok the differences between clusters and standard systems.

As vendors strive for faster processors and denser systems, power and cooling has become a major issue for the HPC market.

Successfully designing a high-performance cluster under today's power & cooling constraints means you need to push the envelope of your system's performance/watt.

In part one, I introduced the two most popular HPC networking technologies -- Ethernet (GigE and 10GigE) and Infiniband. We also compared latency, bandwidth, and the N/2 performance of these technologies. While these numbers give a general feel for performance, there is no easy way to determine the actual performance of your application.

Welcome to the NYCA-HUG

Explicit parallel programming presents special challenges for software developers. Now a new group of languages are coming online to address the compounded problem of multi-core processors on high-performance clusters.

Every time I talk about multi-core, I seem to start out with something like "back in the day" or "when things were much simpler," or some such lament. Now prepare yourself for a stunning bit of insight. Cue music.

HPC cluster optimization is often simple. Avoiding assumptions is hard.

Clusters are getting larger, multi-core adoption plods along, and other findings from our recent HPC micropolls.