Intel Outlines Its HPC Strategy

By Steve Conway

March 16, 2007

Intel has been involved in HPC for more than 15 years. HPCwire asked Stephen Wheat, senior director of Intel's HPC Business Unit, about the company's perspective on HPC as a technology and business market.

HPCwire: What HPC activities is Intel involved in today, beyond supplying processors to HPC systems vendors?

Wheat: Intel is both leading and collaborating on a range of worldwide HPC activities. Aside from developing processors, chipsets and memory subsystems, Intel also has a major platform orientation. Intel works with worldwide Fellow Travelers, which include hardware and software vendors, to deliver and optimize solutions for HPC users. Aside from the hardware platform, Intel makes major investments in software at the applications and middleware levels, including performance analysis tools and an industry-standard MPI implementation that independent software vendors can utilize to maximize their compatibility. Additionally, there is much collaboration with the operating system vendors to ensure the systems software is ready for the latest Intel technology.

HPCwire: Where does HPC fit within Intel's overall technical and market strategy?

Wheat: While Intel has been involved in the HPC market segment for more than 15 years, we have recently formed an HPC business unit. From the standpoint of technical strategy, Intel looks at the HPC market as two segments, high-end and volume HPC. At the high end, there is a technology incubation market for rapid adopters. Beyond that, there is a proven technology segment for the rest of the market. The traditional, high-end segment is much more tolerant of rapid changes and is open to taking advantage of the latest technologies. It can be much more adaptable than segments that are compliance-bound or have a certification process and find it harder to take on new technology. We work directly with the end-user communities in these segments to identify their goals and how we can help achieve them.

In the near term, we look at what product offerings we can add to our lineup, in conjunction with our OEMs and the technology they produce. For Intel, the high-end incubation segment is a place to deploy enabling technology in the near term, and also represents an investment in potential volume technology for the rest of the HPC market and broader production-oriented markets. Incubation is about working to meet the needs of rapid adopters in order to get a larger, more stable group moving to greater performance levels as quickly as possible. We try to help push the innovation downward; and the production, volume-market requirements act as positive constraints on the incubation technology as it trickles down. We're always mindful of what volume segments need as we explore the cutting edge.

HPCwire: Does Intel view HPC as a sector that could provide a worthwhile return-on-investment for the company?

Wheat: Yes. We see the HPC market segment as a growing and viable business opportunity. However, we rarely look at the HPC segment in isolation. HPC innovation quickly migrates into the enterprise segment. There are many opportunities for HPC to influence offerings in the larger markets. Providing a technology pipeline from HPC to the volume enterprise market gives us an opportunity to greatly increase the ROI of our leadership technologies that we first explored and deployed in the HPC market segment.

HPCwire: Is Intel making an R&D investment in HPC, separate from the company's R&D investments in other areas? If so, can you describe the goals for this investment?

Wheat: While we don’t break out our R&D investments, we consider the HPC segment a critical market for technology adoption. We continue to deliver new products and technologies that bring new levels of performance and energy efficiency to this market. An example of this is our new quad-core Intel Xeon processor. Another example is the 80-core terascale research chip that Intel recently announced. While this is not a product, it does demonstrate the type of research and development that Intel is known for.

HPCwire: In five years, what role would Intel want to play within the HPC market?

Wheat: Our goal is to be the platform of choice and deliver technology that brings new levels of performance and energy efficiency for high performance computing. We pursue this goal with a balanced approach to the high-end capability market and the higher-volume capacity market within HPC. Over the next five years, the definition of success would be the clear leadership in HPC market segment share in both the high-end and in the volume part of the market. With the widespread customer acceptance and leadership in performance per watt, price performance, and general applicability to the myriad HPC applications, Intel’s current Core-microarchitecture-based processor line-up, including our currently shipping Quad Core Intel Xeon Processors, is a timely example of such an accomplishment.
 
HPCwire: In an earlier era, the Intel Paragon was the world's fastest Linpack machine. Will Intel ever get back into the business of creating and selling HPC systems?

Wheat: We never say never; however, developing HPC systems that are sold to end users is not our core competency. Having said that, the high-end is requiring increased cooperation between leadership companies to meet the unique challenges posed by the high-end segment. For example, the NSF Cyberinfrastructure program in the USA and the Petascale initiative in Europe each requires a comprehensive approach best achieved by technology collaborations. Therefore we expect to be a very active participant and contributor with our systems customers in the pursuit of opportunities in the high end.

In this way, Intel is best able to support the needs of the HPC market segment more than we had in the past, when we had our own direct product model up through the Paragon platform.

After the Paragon, which was on a non-mainstream i860 Intel processor, Intel developed the ASCI Red machine based on the Intel Pentium Pro Processor, which was our fastest mainstream processor at that time. What we showed there was that industry-standard processors could indeed be the basis of the most powerful computers. We saw several OEMs pursuing similar system designs and concluded it would be best to move forward via a Fellow Traveler strategy. In doing so, we could also support a variety of scale-up and scale-out designs. That's the role we took to increase our influence on the market.

Since then, we have increased the scope of influence, having moved from the basic building blocks model to our platform model where we comprehend how all the ingredients, including software, come together. In the commercial enterprise market, a platform is a server with the relevant software stacks readily available. In the HPC markets, multiple individual systems make up an HPC platform, which then includes the cluster management and provisioning software. Intel participates directly in all of the intermediate levels of integration, but it stops short of marketing a completed system directly to the end-user community.

HPCwire: What are the main things differentiating Intel from other microprocessor companies, where HPC is concerned?

Wheat: First and foremost, Intel's process technology and our ability to apply it to the volume market-at-large, and to make use of this in HPC offerings. No other company has our volume at 65nm, and we will be first with 45nm, then 32nm and so on. That foundation is our greatest differentiation. We can produce high-performance building blocks in a volume that permits us to have very competitive price-performance. The landing zones for trickle down technology are plentiful. There are many opportunities for Intel's high-end technologies to succeed in higher-volume markets. Consequently, our philosophy of innovation and integration will continue to out-pace that of other companies.

Additionally, across the entire ecosystem, Intel brings a level of participation, both driving and supporting platform integration, that is much broader than others. This includes working with the OSVs [Operating System Vendors], IHVs [Independant Hardware Vendors], and ISVs [Independant Software Vendors] at unmatched scales.

HPCwire: Increases in processor clock speed have slowed down significantly. Are there any new technologies on the horizon that could put single-core performance back on a Moore's Law kind of curve?

Wheat: Recent clock speed trends are very real and are due to the broad sensitivity toward power consumption in both enterprise and HPC market segments. The broad market is clearly favoring flops-per-watt over pure flops-per-second. Having said that, we do consider offerings where we increase the clock rate, along with thermal consequences for those who can make use of hotter parts. But in general, performance-per-watt is crucial, and multicores are the market trend for both HPC and the enterprise market.

HPCwire: Multicore processors create new challenges, such as finding enough parallelism within codes and supplying enough bandwidth to each core. As the HPC industry moves to larger numbers of cores, these problems intensify. How does Intel view this issue?

Wheat: Parallelism is an active area of research at Intel. Understanding codes and the mathematics and physics of the problems presents an opportunity to revisit how the algorithms are designed in order to exploit the latest-generation architectures. Application developers typically design for the platforms at hand, and multicore will give them a different platform to deal with. Intel internally has a large investment in the software environment for exploiting multicore platforms. We also work closely with leading academic institutions, standards bodies and our OEMs on the whole ecosystem.

HPCwire: Do you have a many-core (greater than, say, 8 cores) strategy for your x86 and Itanium architectures? If so, will these be part of your HPC strategy? How about the many-core Polaris terascale processor?

Wheat: It's too soon for us to respond to those questions in detail publicly. However, as recently disclosed, the 80-core teraflops research chip demonstrates our technology exploration and research in this area.

HPCwire: With the slowdown in single-core microprocessor progress, a growing variety of custom processors and co-processors are showing up in the HPC market: Cell, SiCortex, FPGAs, GPUs, vector processors, massively multithreaded processors, etc. Many of the world's petascale machine initiatives involve heterogeneous processors. How far will Intel go to embrace this diversity?

Wheat: Intel is about choice at all levels because we do want to be the platform of choice. Heterogeneous computing will be easy to deploy with our PCI Express roadmap. And through the Geneseo program, we are laying the foundation with a large community to ensure that heterogeneous computing is serviced very well. The greater challenge is not so much the connectivity, but the software that allows heterogeneous computing to have continuous benefit. It's important for a platform using an accelerator-of-choice one year to work with next-generation and prior-generation processors. That's a major software challenge. We believe the market will welcome accelerators that make sense and that can support software continuity from one generation to the next.

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