Who’s Driving High Performance Computing?

By Michael Feldman

July 28, 2006

Gone are the days when the U.S. government alone can determine the direction of supercomputing. The commercial growth of HPC over the last two decades has fundamentally changed that dynamic. Adoption of high performance computing in bio-sciences, the financial sector, geo-sciences, engineering and other areas has changed the supercomputing user base in a relatively short period of time.

For many HPC vendors this is a good thing. IDC reports that revenues for the high performance computing market grew by 24 percent in 2005, reaching $9.2 billion. A majority of this revenue is commercial HPC, although the government still represents a significant share. Classified HPC defense spending alone is over a billion dollars.

But maybe more significantly, the vast majority of really high-end supercomputing and cutting-edge research is done with the support of government money. Most multi-million dollar HPC capability systems reside in government-funded supercomputing centers, federal research labs and various undisclosed locations at national security facilities. In the brave new world of commercial HPC, million-dollar-plus capability platforms are the exception. According to IDC, the revenue for these kinds of systems has actually been declining for several years, as less expensive machines have taken their place. Most of the rapid growth in HPC is the result of commodity-based cluster computing, which represented about half of the $9.2 billion in revenue in 2005.

But the U.S. government has some unique problems to solve. Extremely powerful supercomputers are required to support national security applications like nuclear weapons design and testing, cryptography, and aeronautics. Other commercial and scientific applications in areas such as applied physics, biotechnology/genomics, climatological modeling, and engineering can usually (but not always) make due with less-capable systems. Bleeding-edge commercial applications — for example, nanoscale simulation of drug interactions – are emerging, but most of these are being facilitated by government support.

A March 2006 report by the Joint U.S Defense Science Board and the UK Defence Scientific Advisory Council Task Force on Defense Critical Technologies concluded the following:

“Multiple studies, such as the recently completed [November 2004] National Research Council study, conclude that '…the supercomputing needs of the government will not be satisfied by systems developed to meet the demands of the broader commercial market.' The government must bear primary responsibility for ensuring that it has the access to the custom systems that it requires. While leveraging developments in the commercial computing marketplace will satisfy many needs, the government must routinely plan for developing what the commercial marketplace will not, and it must budget the necessary funds.”

Last week at a hearing before the Senate Subcommittee on Technology, Innovation and Competitiveness, several industry and government representatives offered testimony to address some of these issues.

One of the industry representatives to testify was Christopher Jehn, vice president of Government Programs of Cray Inc. He sounded that alarm that advances in HPC technology have slowed and that the promise of commodity-based supercomputers has not materialized. He attributes this to the fact that general-purpose processors and other commodity-based technologies used to build supercomputers were designed for other purposes — essentially, personal computing and enterprise computing. The result is that scientists must expend a lot of effort to get HPC software to run efficiently on these homogeneous commodity-based machines.

“Over the last decade, the computer industry has standardized on commodity processors,” observed Jehn. “With high volume low-cost processors, supercomputer clusters consisting of commodity parts held out a promise to users of ever-more powerful supercomputers at much lower cost. At the same time, the federal government dramatically reduced investments in supercomputing innovation, leaving the future of supercomputing in the hands of industry. But from industry's perspective, the supercomputing market is not large enough to justify significant investment in unique processor designs and custom interconnects — as the supercomputer market is less than two percent of the overall server marketplace, according to International Data Corporation. To advance supercomputing, industry has relied on leveraging innovation from the personal computer and server markets.”

This reflects Cray's “Adaptive Computing” pitch — to advance HPC in a meaningful way, we have to move from homogeneous systems to heterogeneous ones. Just scaling up the current architectures won't get us there. The implication is that the government needs to make a significant investment in technology beyond commodity-based computing.

Dr. Irving Wladawsky-Berger, vice president of Technical Strategy and Innovation, IBM, offered a different perspective. He warned that the government cannot afford to ignore market realities when funding HPC projects. During his career he witnessed the failure of supercomputing companies that relied solely on government-based projects and were heedless of marketplace requirements. He related IBM's success with its Blue Gene architecture as an example of leveraging commodity technology — in this case, PowerPC processors — to build cutting-edge systems.

“Supercomputing was once confined to a niche market, because the hardware was so very expensive,” stated  Wladawsky-Berger. “That changed over time with the introduction of workstation and PC-based technologies, the latter becoming immensely popular in Linux clusters during the late 1990s. Today, we even use low-power, low-cost micros — consumer-based technologies — to attain very high degrees of parallelism and performance, as in our Blue Gene system, which has reached a peak of 360 trillion calculations per second. Now, we are seeking to build supercomputers using technologies from the gaming world, such as the Cell processor. All these approaches leverage components from high-volume markets, and aggregate them using specialized architectures; thus the costs are significantly lower than in earlier days and the potential markets are consequently much bigger.”

From IBM's point of view, Blue Gene is an affirmation that commodity-based supercomputing is a practical model for the future and the government should pay attention to market viability as it looks to invest in new programs.

Dr. Joseph Lombardo, Director National Supercomputing Center for Energy and the Environment at the University of Nevada, Las Vegas, described some of the history of the U.S. government's past investments in high performance computing. He suggests that our federal HPC interests, academia and the larger HPC community are all intertwined and the government needs to act accordingly. He noted that after a brief period of interest in “Grand Challenge” applications in the late 1980's and early 1990's, the government switched its focus to distributed computing and COTS technology. He said while these initiatives led to a broader range of individuals working in scientific computing, it also resulted in starving the high-end of HPC R&D. But after the rise of Japanese supercomputing in the 1990's, the U.S. government once again refocused its efforts in high-end supercomputing

“At the end of the 1990's DARPA and other organizations began to see that foreign countries, such as Asian groups, were overtaking the U.S. position in high performance computing once again, and recommended policies that would fund and support the high end of the field once again,” said Lombardo. “The DARPA High Productivity Computing Systems program is a good example of this shift back toward an emphasis on high-end capability. The DARPA program is focused on providing a new generation of economically viable high productivity computing systems for the national security and industrial user community in the 2010 timeframe. This trend has continued with the High Performance Computing Revitalization Act, the President's 2006 state of the Union Address, and with the FY 07 budget which increased DOE's high performance computing programs by almost $100 million.”

Lombardo's comments suggest that we can balance the government's and industy's need for advanced supercomputing with market realities. He points to the DARPA HPCS program as an example of this approach.

But HPCS may also expose a potential conflict in the government's role. The program's stated goal of “providing a new generation of economically viable high productivity computing systems for national security and for the industrial user community” suggests that HPCS intends to address both the government's and industry's supercomputing needs, and do so within a commercially viable framework. The implication is that all these objectives are compatible.

But national security represents a rather specific set of very high-end supercomputing applications, while the industrial user community represents a very diverse range of HPC users. Can a single supercomputing model (or two) satisfy everyone? Even if we limit the industrial users to potential petascale customers like Boeing, I might still ask the same question.

And what is really meant by “commercially viable?” For supercomputing systems that push the envelope, commercial viability has always been problematic. Vendors usually don't expect such systems to make money straight out of the lab. I understand the desire to produce a general-purpose petascale solution, but I guess it makes me uncomfortable to think the government is going to try to predict the economic viability of a future architecture. After all, DARPA isn't a market research firm.

As commercial HPC continues to expand, the government will be increasingly challenged to control the direction of future supercomputing architectures. Market realities are pushing the hardware and software in a different direction than the needs of some critical high-end HPC users and will probably continue to do so. Such is the nature of capitalism, which, like processor scalability, has its limits. Market forces don't automatically produce optimal results. The government role in HPC, as in other areas, should be to support our national interests.

—–

To find out more about what took place at the Senate's Subcommittee on Technology, Innovation, and Competitiveness hearing on HPC, take a look at our feature article that describes these proceedings. To learn more about the evolving relationship between HPC, the government and business competitiveness, read this week's interview with Suzy Tichenor, Council on Competitiveness vice president, and Bob Graybill, senior Council advisor.

As always, comments about HPCwire are welcomed and encouraged. Write to me, Michael Feldman, at [email protected].

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire