Ties That Bind: Organizing Large-Scale HPC in the European Union

By John E. West

June 19, 2008

The combined economic resources of the 27 member states of the European Union make it the largest economy in the world. No stranger to the world of HPC, and convinced of the power of technology to further propel the prosperity and well-being of its citizens, the EU is investing heavily in pan-European resources that it hopes will propel its members to the forefront of the computational stage.

Viviane Reding, the European commissioner for Information Society and Media, believes in the power of technology to transform lifestyles and drive the prosperity of the European Union. Her portfolio stretches from telecommunications to eHealth services. And she believes in HPC.

A recent Computerworld article captured her remarks at the opening of the Ter@tec 2008 HPC conference in France, “Supercomputers are the ‘cathedrals’ of modern science, essential tools to push forward the frontiers of research at the service of Europe’s prosperity and growth.” But, as many of us know firsthand, building a large-scale supercomputing infrastructure to support research and industrial goals is complicated and can take enormous financial resources.

The computers themselves, often costing upwards of $100M for leadership resources, are just the down payment. Indeed, modern HPC center planners now count on the lifecycle costs of supercomputers to outstrip their acquisition costs.

Users must be able to connect to supercomputing centers over high bandwidth, low latency, very wide area networks, and they must be supported through routine questions and complicated software issues. The systems require continuous care and feeding by trained administrators. Multi-thousand ton chillers must be bought and installed, and building power infrastructures must be expanded to support the multi-megawatt needs of modern supercomputers.

As the two largest economies in the world, the United States and Japan have been in a position to devote the resources needed to develop and sustain world-leading supercomputing infrastructures. While each of the countries of Europe have made, to varying degrees, national investments in HPC, those economies individually simply didn’t have the resources to compete with the US and Japan. The collective resources of the European Union, however, may dramatically alter that dynamic.

Collectively the 27 member states of the European Union represent the world’s largest economy. Through its various science and industrial research programs, the EU, along with individual members acting in concert, has committed to creating a supranational supercomputing infrastructure on a par with anything the US and Japan have created to date.

One of the key focal points for HPC in the EU is DEISA, the Distributed European Infrastructure for Supercomputing Applications. DEISA is a consortium of national supercomputing centers in Europe, brought together with a common infrastructure to advance broader European computational science research goals. I spoke with Stefan Heinzel, the project coordinator for DEISA, and Hermann Lederer, who leads external relations, by email about the effort.

I asked them to describe how DEISA fits into the science ecosystem in Europe. “An agency like the US National Science Foundation is a national science policy and funding agency, as are research units of the European Commission (EC) at European level, and respective national organisations,” says Lederer. “The EC is co-funding the DEISA2 project along with various national partners, as the NSF is funding the TeraGrid project. DEISA is an HPC infrastructure project and can be regarded as a European counterpart of TeraGrid.”

There are 11 principal partner centers from seven European countries participating in DEISA, including the Rechenzentrum Garching of the Max Planck Society (RZG), the Barcelona Supercomputing Center (BSC), the Jülich Supercomputing Centre (JSC), the Edinburgh Parallel Computing Centre (EPCC), and others. The partnership has been operational for some time, having just received a new round of funding from the Seventh Framework Programme Research and Technological Development, the EU’s research funding arm.

As explained by Lederer, the DEISA infrastructure is layered on top of existing national supercomputing centers, each with its own skills, funding, and computational goals. Each of these national resources sets aside a portion of its resources for computational projects at the European level. No hardware funding is awarded through DEISA; all upgrades to HPC systems are funded by individual member states and institutions. Starting with only 30 teraflops in 2004, DEISA sites have grown to roughly 1 petaflops of aggregate performance in 2008.

DEISA partners are connected to users, and to one another, via GEANT2 and the National Research and Education Networks. This infrastructure connects DEISA sites at 10 Gbps, with dedicated wavelengths on the NRENs and GEANT2.

DEISA encourages researchers to stretch the boundaries of their understanding on important research issues through the DEISA Extreme Computing Initiative (DECI). DECI projects are awarded large blocks of time on DEISA HPC resources to tackle especially challenging, large-scale computational problems. In DEISA1, DECI awards were made on individual projects. Beginning this year with DEISA2, the EU is shifting focus to projects that will support Virtual Science Communities across Europe, strengthening the overall R&D environment and encouraging broader collaboration. DEISA is also shifting the technologies and expertise it invests in to provide tools that will integrate broader teams, including resource integration via grid technologies, and Web tools and portals. Interestingly DEISA provides data management services and an infrastructure for the European research community that includes a common multi-cluster filesystem, accessible from nearly any DEISA HPC platform, based on IBM’s Multicluster GPFS.

While the 1 petaflops of aggregate computational power deployed by DEISA partners is certainly significant, the EU and European HPC leaders have recognized a need to further increase their investment to support goals of advancing the prosperity of Europe through global computational leadership. Enter PRACE, the Partnership for Advanced Computing in Europe.

The goal of PRACE is to establish three to five European “tier 0” centers, each with petascale resources, that will serve broader EU science and industrial research goals. PRACE is still very much a planning exercise, currently funded at 40M euros of what is expected to grow to an estimated 200M euro budget for operations alone. PRACE envisions a three-year lifecycle for its large machines, and doesn’t intend to operate them in isolation. According to Lederer, “PRACE and DEISA activities are expected to merge for the operation of tier 1 and future tier 0 centers in an integrated European HPC ecosystem. Most PRACE members are already DEISA partners.”

In contrast to DEISA, which is a bottom-up approach organizing existing centers, PRACE is a top-down exercise. Why the shift? I talked with Achim Bachem, coordinator of the PRACE project, by email. “Combining national resources in a bottom-up way merges capacities, but does not provide larger capabilities. Procuring systems of the highest performance level requires the combined financial efforts of more than one country,” says Bachem. “The second reason is that we want to create a homogeneous pan-European HPC service with common access policies and a single peer review system. This will be an important contribution to the European Research Area and also requires a top-down approach.”

PRACE just started operations in January of this year, and they have a lot of work ahead of them. Their first challenge is to determine what legal entity is best to conduct PRACE business, and how it will be organized. According to Dr. Bachem, this won’t be easy, “This is an essential and complex task, since national as well as European interests have to be combined into a single transparent and efficient framework.”

While things get organized, progress is being made on the technical front with a survey of architectures and candidate petaflops systems for 2010. “Based on the suitability of the various architectures for important applications, we have defined a set of prototype systems that will cover all relevant architectures, ranging from well proven ones like MPPs to upcoming hybrid systems,” says Bachem. “Since this joint proposal has an overall budget of several million euros, it will be evaluated and approved by the European Commission in July, before we start the acquisition.”

Will the focus be on production-quality systems or novel architectures? According to Dr. Bachem, both: “…one of the advantages of this joint European venture is that we will be able to operate more than one leadership system at any given point in time. This gives us the freedom to also invest in more advanced technologies.”

PRACE isn’t only focused on hardware. Its largest work package (in person-months) is devoted to software for petascale systems. “In the past,” says Bachem, “software has been a European strength, but currently there is nothing comparable to the SciDAC initiative in the US. Europe has to be more active in this area in the future.”

This effort is clearly a major potential benefit for the European (and global) scientific and research community, but there is significant time pressure to deliver results quickly. Implementation is not scheduled to begin until late 2009 or 2010, with much of the next two years focused on organization and planning. With the release of the latest TOP500 list officially marking the passage of the petaflops hurdle this week, petascale systems, while certainly not routine by 2010-2011, won’t be major scientific accomplishments either. In order to ensure the relevance of this substantial program, the EU knows it has to act as quickly as it can.

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