HPCwire: HP Launches "Catalysts for HPC Innovation" Initiative

The Leading Source for Global News and Information Covering the Ecosystem of High Productivity Computing

Leading HPC
Solution Providers

HPCwire >> Off the Wire

November 02, 2007

HP Launches "Catalysts for HPC Innovation" Initiative

Driven by research and development, new High Performance Computing (HPC) technologies are continually evolving, enabling scientists and researchers to solve larger and more complicated problems. HP, an industry leader in HPC, is well known for its extensive history of innovation and promoting initiatives that pursue emerging technologies.

To leverage its customers' insights and BT challenges, HP recently launched a new emerging technology initiative called, "Catalysts for HPC Innovation." Scott Misage, director of advanced development programs and strategic customer management for HP's High Performance Computing Division, shares some of the details:

Question: What exactly is the Catalysts for HPC Innovation initiative?

Scott Misage: It's a program designed to increase HPC innovation in a customer focused way. The 'Catalyst' initiative includes a set of advanced development projects and customer collaborations that provides HP the opportunity to evaluate and drive adoption for the next wave of emerging high performance technologies.

Our goal is to closely collaborate with our customers to incubate new technologies in real-life situations, covering a range of technical areas. The program provides early insight into future industry needs and shortens the innovation cycle, so that our new business technology solutions anticipate and solve real customer problems quickly.

Question: HPC vendors have always put a tremendous amount of time, money, and effort into research and development. What makes this initiative different from what has been done in the past and what other companies are currently doing?

SM: It's the upfront customer involvement that makes all the difference. We are not taking an ivory tower approach to these projects. HP's program is unique in that we collaborate with our customers to identify new opportunities, to drive the requirements and to provide us with advice during the design phase. Without the customer interaction, we would be missing the opportunity to better conceptualize and integrate features that solve their real problems.

Question: What are some emerging technologies that fall under the initiative?

SM: We have projects that encompass computation, data management and visualization. In the computation area we are investigating multi-core optimization, accelerator and dense computing technologies. Data management projects include converged fabrics, remote caching and grid-enabled shared file systems. In the visualization area, we are exploring things like parallel compositing. Of course, we are also working to extend our leadership technology in areas that address environmental concerns, such as power and cooling, in other words, addressing customer concerns for "green" buildings and work systems.

Question: Tell us about the Multi-core optimization project.

SM: We're entering a phase where taking full advantage of the power of multi-core processors is critical for customers to accelerate innovation and to improve their business success. HP's Multi-Core Optimization Program, announced in June, is a broad initiative to optimize HPC applications on multi-core systems and enable customers to realize the full performance benefits of multi-core technology. Like all the 'Catalyst' projects, our multi-core R&D program features a unique collaborative approach that combines HP developments with those from a broad set of technology leaders, including the Karlsruhe Institute of Technology (KIT), Advanced Micro Devices (AMD) and TotalView Technologies. We are working with these customers and partners to provide open, non-proprietary solutions that enhance multi-core performance across a variety of HPC architectures, platforms and operating environments.

Question: Accelerator technology is an area that has received a tremendous amount of press in the HPC community, so it is apparent why one would be on your list. Can you tell us what the Catalysts for Innovation initiative is working on with this emerging technology?

SM: We recognize that many vendors and users are proposing alternative technologies such as General Purpose Graphical Processing Units (GPGPUs), Field Programmable Gate Arrays (FPGAs) and custom ASICs that will deliver substantial increases in application performance on industry standard platforms.

Our accelerator project is an on-going effort to research various acceleration technologies for HPC applications and benchmark them to determine which ones are most beneficial for HPC customers. We are also collaborating with software tool vendors to help ensure their products work well in the HP HPC environment.

Question: Can you give us a specific example of a customer you are collaborating with on accelerator technologies and how they are benefiting?

SM: HP and Intel are working with Vital-IT (http://www.vital-it.ch) to investigate the use and advantages of different accelerator technologies in the life science area. We are looking at enhancing existing codes or libraries, such as "R" a major research tool in the field of statistical computing. The research collaboration also includes new applications using heavy simulations of public domain software in the field of cancer research and system modeling.

By working together with HP on the Catalysts program, Vital-IT can allow these computing technologies to be used by their researchers now. Vital-IT worked directly with the engineers at HP and our partners, to jointly understand their specific issues and challenges. Vital-IT also benefited from exposure to other HPC participants in our program, enabling them to address common issues and problems as well as promoting inter-customer interaction.

Contrast this with an approach where someone invents something, keeps it a big secret until it is "done," then takes it to customers and says: you need to change the whole way you work to take advantage of this great thing. Our hope is to gradually enhance the accessibility of these rather "niche" technologies to a wider researchers audience. Now most of the interest is targeted towards systems biology, which in essence is large biology with big computational need.

Question: Converged fabric technology is also a hot topic in HPC. Could you give us some details about this project?

SM: Converged fabrics address the need to simplify datacenter management by consolidating all communications, such as LAN, SAN and cluster, onto a single fabric. We expect this architectural approach to significantly lower infrastructure and management costs without sacrificing performance. In order to better understand the characteristics and analyze the behavior of the datacenter with converged fabric technology, HP has established a converged fabrics lab where we can demonstrate results with real customer workloads.

Question: You mentioned parallel compositing as a technology you are developing for advanced visualization. Tell us more about that and the customer collaborations in that area.

SM: HP's open source Parallel Compositing Library and API is groundbreaking in that it greatly simplifies the development and use of parallel applications on graphics clusters. Announced in May, the Library and API is a tuned, documented and supported tool to help customers take further advantage of open source and Linux in their datacenters. The API was developed by HP over the last several years with input from industry thought leaders such as Computational Engineering International (CEI), CSCS, BME (Budapest University of Technology and Economics) and the University of Manchester, U.K.

The Library and API can do for graphics clusters what the MPI did for compute clusters -- allow customers to take advantage of the inherent performance scalability of clusters with network-based pixel compositing. The huge data sets created by complex simulations and models can now be interactively rendered and visualized. This helps scientists and engineers better understand the results of applications such as seismic analysis, engineering design and simulation, medical imaging, weather models and animation that demand more performance than a single graphics system can deliver.

Ongoing collaborative work includes evolving visualization algorithms and developing new algorithms that take advantage of the parallel computational and rendering capabilities of a graphics cluster. This work will allow visualization of large data sets and complex processes that otherwise could not be performed interactively. HP recently demonstrated BME's work at LinuxWorld on an HP Scalable Visualization Array (SVA.)

Question: Customer collaboration in HPC R&D makes a lot of sense. Has HP ever done anything like this before?

SM: Actually, the idea of doing advanced development work collaboratively with customers is nothing new for HP. We've been doing it for years in the HPC space, most notably with US Federal customers, such as the Department of Energy (DOE). We've had some pretty significant successes, which is why we are confident that collaboration with customers is the way to go.

Question: How would you compare the Catalysts program from work being done at HP Labs?

SM: There is no "black and white" line between the work at HP Labs and the work in our division. We work closely on a number of projects to bring innovative new technologies; solving real problems. HP Labs' function is to deliver breakthrough technologies and technology advancements that provide a competitive advantage for HP, and to create business opportunities that go beyond HP's current strategies.

For example, Labs is working on nano-photonic technologies -- essentially a nanotechnology connecting silicon to silicon with light. That's groundbreaking technology with potentially far reaching application across many platforms and useful in everything from handheld devices to supercomputers. Researchers at HP Labs are charged with inventing for the company's future. This can entail technology solutions that might be 3-5 years from commercialization, whereas our division tends to be more narrowly focused on "applied" research in the form of finished products.

Question: We have seen how HPC is frequently at the forefront of key trends and a market leader in terms of creating and evolving disruptive technologies. Do you expect these emerging technologies to have substantial impact beyond enterprise computing markets?

SM: Yes, we do. The mission and behavior of HPC customers frequently places our market at the forefront of key market trends--our customers demand and reward innovation. Several of today's hottest market trends, such as scale-out computing and Linux, were evident years ago in HPC. HPC also pioneered the use of AMD processors as a viable alternative to Intel products, and our market is leading the adoption of bladed infrastructures.

All customers demand innovation to solve new problems. We know investments in this space return dividends in the fast growing HPC market, but they will also eventually make their way into the commercial space.

Question: What's in it for HP? How does HP benefit from this program?

SM: We recognize that customer collaboration provides better opportunities to develop innovative products that make a difference in real customer environments. The collaboration helps us to execute on our value proposition -- giving us a tremendous competitive advantage in meeting customers' needs and exceeding their expectations.

Scott Misage serves as the Director of Research and Advanced Development in the High Performance Computing Division of HP and has more than nine years of experience in various HPC roles. In his current position, Misage identifies and manages projects aimed at bringing innovative new technologies to bear on future high performance computing challenges. Prior to HP, Misage held various information technology positions in the U.S. Federal Government, Central Intelligence Agency and the U.S. Department of State.

Article Tools

Share & Save Options

Discussion

There are 0 discussion items posted.

Join the Discussion

You must be a registered user in order to comment on HPCWire stories.

Members enjoy:

	Access to Special Content, including Podcasts, Webcasts, and Whitepapers and more!
	Ability to take an active role in the discussions that are shaping the HPC technology discipline!
	Special offers available only to HPCWire members!

Become a Registered User Today!

Registered Users Log in Here to Comment

Email Address:	Password (case sensitive)

Remember me	Forgot Password?

RSS Feeds

Feeds by Topic

Feeds by Industry

Feeds by Content Type

Feature Articles

Sun, Fujitsu Deliver Quad-Core SPARC64 Servers

Those of you looking forward to Rock -- Sun's much anticipated 16-core processor originally scheduled for release later this year but now pushed to the second half of 2009 -- don't have to wait for those chips to come out to experience that launch party euphoria. This week Sun and Fujitsu announced the latest of their enterprise line of SPARC-based servers, sporting the new SPARC64 VII chip.
Read More...

The Week in Review

The UK makes a multi-million pound investment in science and computing; the Defense Department funds a HPC software project; and TACC's Ranger shows off its new Opterons. John West recaps those stories and more in our weekly wrap-up.
Read More...

Sun's Fortress Language: Parallelism by Default

If anyone knows how to introduce a new programming language, it's Sun Microsystems. The company's highly successful Java language, which was introduced in 1991, has become ubiquitous in network-centric and embedded computing. Today, there's a whole research team at Sun Labs devoted to programming languages, and the big project there in recent years has been the development of the Fortress programming language. The end game is to "do for Fortran what Java did for C."
Read More...

Read more HPCwire features...

Top Headlines

Biology Enters 'The Matrix' Through New Computer Language

Jul 22 | Harvard Medical School | A team of Harvard Medical School researchers have developed a computer programming language that can be used to model the biomolecular behavior of proteins. Read more...

Nvidia: Larrabee is a Reaction to CUDA

Jul 21 | Custom PC | Nvidia responds to Pat Gelsinger’s comments about CUDA being just a ‘footnote’ in computing history. Read more...

PlayStation Processor Dominates Green Supercomputing

Jul 21 | ElectronicsWeekly.com | Computers based on the Cell processor dominate the world ranking for energy efficient supercomputers, according to the just-published Green500 list. Read more...

More Power7 Details Emerge, Thanks to Blue Waters Super

Jul 21 | IT Jungle | Rumors have been circulating about IBM's future Power7 processor and how the chip fits into NCSA's upcoming "Blue Waters" supercomputer. Read more...

Hello AMD Socket G34

Jul 17 | DailyTech | AMD's 12-core and 8-core processors will get a new home in 2010. Read more...

Read more headlines...

Featured Whitepapers

Improving Performance and Manageability for Seismic Processing and Imaging Applications with Parallel Storage

Jun 05 | | As pressure increases on the upstream seismic processing community to deliver ever-higher levels of productivity and efficiency, a new generation of storage solutions will be required that allow the maximum utilisation of high-performance computing (HPC) Linux cluster resources, together with the minimum of management overhead.

View the White Paper Library

Multimedia

Podcast: Interview with Ben Bennett of ClearSpeed Technology

Today, HPC organizations are requiring substantially more floating point performance to solve real-world problems. In this podcast, Ben Bennett, ClearSpeed General Manager, discusses how acceleration technology can improve the overall performance of standard x86-based systems...