INCITE Program Targets American Competitiveness

By Steve Conway

February 9, 2007

In this exclusive HPCwire interview, Dr. Raymond Orbach, director of the Department of Energy's (DOE) Office of Science and the nation's first Under Secretary for Science, and Council on Competitiveness President Deborah L. Wince-Smith discuss the value of DOE's INCITE program. INCITE awards huge blocks of time on DOE supercomputers to university, industrial and national laboratory research projects with strong potential for HPC-assisted breakthroughs.

HPCwire: How and when did the INCITE program come about?

Dr. Orbach: We launched the program in 2003, more out of instinct than anything else. We believed science and industry would profit from greater access to high-end computation. The problem back then was that in the U.S., we in the government were awarding computer time more to maximize the number of users, with the result that everyone got a little time. This proved not to be an efficient way to address real problems. I know from my own experience what it's like to work on challenging problems when your computer time is limited.

We asked ourselves how discovery could advance if we gave people enough time to solve major problems and allocated the machine time based on peer review. We started with just four proposals, but when people found out we'd be allocating as much as a million hours or more per project, INCITE really caught on.

HPCwire: Has the program turned out as expected? Have there been any surprises or learning experiences along the way?

Dr. Orbach: INCITE has proven itself beyond any expectations we had. For 2007, nine industries were awarded time. One lesson from the program is that researchers in industry are fully on a par with researchers in universities and can have problems that are just as challenging. This was a surprise. Another surprise is that computational speeds have increased so rapidly. In 2003 when the program started, I was skeptical that we would have a petaflop machine any time soon. Now we have a shot at that at Oak Ridge by the end of the next fiscal year. Petaflop computing will make the INCITE concept even more powerful. We'll be able to do convergence on Navier-Stokes equations, for example, which I thought we'd never be able to do. The opportunities are quite remarkable.

HPCwire: What led the Council to recommend that DOE extend the program to U.S. industry in 2005?

Ms. Wince-Smith: The Council's fundamental belief is that U.S. competitiveness and the nation's ability to add high-value economic activity increasingly depend on 21st-century modeling and simulation. DOE has been a global leader in using high-end HPC systems for government missions requiring capabilities at the frontier of computing. As U.S. taxpayers, we have all invested in these HPC capabilities. The INCITE program is leveraging these investments not only to advance the nation's scientific leadership, but our industrial competitiveness and standard of living.

HPCwire: Why did you choose to collaborate with the Council for the INCITE program?

Dr. Orbach: The Council on Competitiveness brings a wealth of partners from industry that DOE would not otherwise be able to reach. We chose to collaborate with the Council in order to broaden the reach of INCITE, and that choice has proven itself time and again to be the right one. Like DOE, the Council and its members care deeply about American competitiveness. Collaboration with the Council — particularly in the area of high performance computing — is a natural extension of our mutual interest in American leadership in basic and applied research.

One problem government has is choosing winners. People get very upset if they think this might be happening. We wanted to team up with an organization, in this case a private organization, that would be open to the private sector and acceptable to the private sector as a vehicle for transmitting proposals. The Council also recognized early on the advantages that high-end computing provides to industry. They knew it could reduce time-to-market and prototyping costs if industry could get large blocks of computer time and some help with software. In our companion program, SciDAC, DOE works with science and industry to improve their codes. The program brings in mathematicians, scientists, computer scientists and others to work as a team. The sociology of computing has really changed. There are now people from different fields collaborating as teams to optimize codes. This is what makes the U.S. special. Other countries are not doing this to anything like the extent we are.

HPCwire: Dr. Orbach talked about the growing importance of multidisciplinary HPC research. What's the Council's view on this?

Ms. Wince-Smith: This ties back to the Council's view on 21st-century innovation, and how different today's research is from the way research was done 10 to 20 years ago, when the primary approach was individuals working in centralized research facilities and in sharply demarcated disciplines. Today's biggest insights are coming from the fusion of knowledge from different fields, and from applying HPC to these multidisciplinary problems. For studying disease pathways through the body, for example, knowledge is needed about physics, chemistry, biology and in some cases also nanotechnology. HPC is going to accelerate the demonstration and implementation of multidisciplinary science, including within business and industrial settings.

HPCwire: What are the benefits of the INCITE program for U.S. industry?

Ms. Wince-Smith: We see them as three-fold. First, INCITE provides an opportunity to tackle hard, complex problems that otherwise can't be investigated well today. In this sense, INCITE extends the R&D capabilities of program participants. A second important benefit is that INCITE participants from industry form new relationships with individuals at national labs who understand how to exploit cutting-edge HPC architectures and simulation on very advanced problems. Finally, INCITE gives industrial participants a competitive advantage in moving to next-generation problems and servers. The INCITE work helps them makes decisions about which way to go with their product development. That can be worth millions of dollars to the companies.

HPCwire: What are the benefits of industry participation for DOE?

Dr. Orbach: This program is about competitiveness. We are part of the President's American Competitiveness Initiative, and in his State of the Union address, this year and last, he talked about the importance of basic research to the competitiveness of our country. At DOE, we're using our computing resources to enhance the competitiveness of U.S. industry. At a recent Council board meeting, I heard story after story about offshoring and the competitiveness of other countries based on labor costs. I talked about something that's uniquely American: having these high-end machines. We've had many meetings and workshops, and the Council's done multiple studies, and we know what's needed. We need more virtual prototyping. Pratt & Whitney, Procter & Gamble, GE, Wal-Mart and other major companies told us they could save tens of millions of dollars if they could simulate their most challenging problems. Boeing has substantially reduced the number of physical prototypes they build and the number of wind tunnel tests they conduct. Physical experiments are very expensive and time-consuming. These companies and others have gotten confident enough in computer simulation and modeling to cut back on the physical experiments, and this is making them more competitive.

HPCwire: Are there empirical metrics for evaluating the success of the INCITE program? For example, do participants need to submit periodic reports to the DOE?

Dr. Orbach: We get reports every year. Keep in mind, however, that working on computer architectures at this scale is new for most of the INCITE awardees. These companies had access before to maybe 5,000 or 100,000 CPU hours.

This is a learning process for all of us. The first lesson is that you often can't do everything in one year. It takes a while to optimize codes on these architectures. That's why we've given some two- and three-year awards. We get a report from every INCITE awardee each year that we have peer reviewed. Our ultimate metric, though, is discovery. The real question is, are we able to enhance scientific and industrial discovery with the help of high-end simulation?

HPCwire: What kind of feedback has the DOE gotten from scientific and industrial participants in the INCITE program?

Dr. Orbach: The feedback has been overwhelmingly positive. [See comments at end of interview.] It indicates that the community thinks this program is for real. Another indication is that the number of proposals has shot up quickly. For 2007, we got requests for twice as much time as we have available, even though we substantially increased the number of available CPU hours. The requests totaled 180 million CPU hours, versus the 95 million we were able to award. We have a variety of architectures in the INCITE program. NERSC has been oriented more toward capacity machines to handle requests for smaller numbers of CPU hours, while at Oak Ridge and Argonne we have capability machines where we allocate large blocks of time to a more limited number of users.

HPCwire: Do the 95 million hours represent excess capacity on these computers, which would have gone unutilized if not for the program?

Dr. Orbach: No. These hours represent an opportunity we created by investing in these architectures. We didn't know at the start what the efficiency of each of these architectures would be. We're learning about that in practice and finding out which architectures are most effective for which problems. Our high-end machines have always been oversubscribed, but our belief from the beginning has been that making them available through the INCITE program is what's needed to advance science and industry.

HPCwire: In sum, what do you think the INCITE program does for U.S. competitiveness?

Ms. Wince-Smith: It propels us into performing the high-value economic activity on which our prosperity depends. The U.S. can no longer compete successfully on low-end, commoditized engineering work. We need to push the innovation envelope and compete from the frontiers of these disciplines and fields. INCITE isn't just about getting access to these powerful computing tools; it's even more about what they enable researchers to think about. This program helps expand the horizons of scientific and industrial thinking, and this leads to greater competitiveness.

One area with great potential, for example, is the intersection between our current energy portfolio and moving away from carbon and toward sustainability. This in itself is a crucial, fascinating problem that will be addressed both in the national labs and by industry researchers. To remain competitive, we need to keep creating the future. Supercomputing gives us the ability to do this.

INCITE is truly a great example of a public-private partnership. It takes real leadership to do in the public sector what Dr. Orbach and his team have done with this program. They had the foresight and conviction to recognize and grasp these large opportunities. We at the Council couldn't have gone forward with our HPC Initiative without a public sector partner like Dr. Orbach. I think he will have a profound legacy, and we're honored to work with him and DOE on the INCITE program.

HPCwire: I would ask you the same question, with respect to scientific and industrial competitiveness.

Dr. Orbach: The question remains to be answered over time, but we can already see some important progress. New findings on the rotation of neutron stars by an Oak Ridge researcher were reported in the January 2007 issue of Nature magazine and considered for that issue's front cover. These findings would not have happened, at least not nearly as soon, if Tony Mezzacappa had not had access to these high-end computing resources. This is what I mean by discovery. Simulation is really proving itself as the third pillar of scientific discovery, fully on a par with theory and experimentation. We have enough confidence in simulation that we are continuing to enhance the speeds and variety of the architectures we use. Sandia co-developed the large “Red Storm” system that became the basis for the Cray XT3 at Oak Ridge. The Blue Gene/L at Argonne will morph into the Blue Gene/P. NERSC will move into new architectures and split between capacity and capability machines. My personal belief is that scientific discovery can advance well in this way.

—-

For comments about the program from some of the INCITE participants, visit http://www.hpcwire.com/hpc/1255549.html.

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