INTERVIEW: EuroHPC Director Kalbe on Plans to Ramp Up Europe’s HPC Infrastructure

By Steve Conway

January 29, 2019

Editor’s Note: November 6, 2018 marked the first meeting of the governing board of the European High Performance Computing Joint Undertaking (EuroHPC), a legal and funding entity which will enable pooling of EU and national resources in HPC with the aim of developing a pan-European supercomputing ecosystem and providing access to leadership-class supercomputers developed and deployed in the EU. In this exclusive for HPCwire, Steve Conway, Hyperion Research senior vice president of research, interviewed Gustav Kalbe, head of the High Performance Computing and Quantum Technologies unit of the European Commission. Dr. Kalbe is also the interim executive director of EuroHPC.

HPCwire: Can you summarize what the European Union is doing in HPC, especially the EuroHPC Joint Undertaking?

Gustav Kalbe, interim executive director, EuroHPC

Dr. Kalbe: EuroHPC is a legal entity with a mission to develop European technologies for the coming exascale era.  EuroHPC combines funding from the European Union with funding from European states participating in the EuroHPC initiative. We want to make a European HPC infrastructure available that’s competitive with the best in the world, first by acquiring next-generation supercomputers and putting them into service for user applications by 2020, and second by supporting research and innovation activities to deploy exascale machines in 2022 so users can continue to benefit from world-leadinng HPC technologies.

HPCwire: What is your role as interim executive director of EuroHPC?

Dr. Kalbe: EuroHPC was established about in November 2018 but as an empty shell with no staff.  My role is to find a building to house EuroHPC, hire an executive director and staff, establish business processes and so forth, so that by late 2019 EuroHPC is really moving along well.  As interim executive director, I have the authority to sign contracts, launch calls for pre-exascale and exascale supercomputers, and carry out the functions that need to happen until a permanent executive director is in place. The EuroHPC governing board makes the final decisions, but I can commit budgets and launch activities.

HPCwire: How is the European approach to HPC different from the strategies in the US, China and Japan?

Dr. Kalbe: The biggest difference is that we aren’t just one sovereign state. The European Commission and nations participating in EuroHPC have to pool our funding and our efforts to create globally competitive HPC systems and services that will exist as a shared infrastructure that users from all the participating countries have an equal right to access.  Another difference, at least compared to some other areas of the world, is that Europe today is not making some of the components needed to build globally competitive European supercomputer systems. We rely a lot on key technologies and components from elsewhere in the world. That’s something we’re trying to remedy by 2022-2023, so we can produce globally competitive European HPC systems that are based on indigenous European technologies and components.

HPCwire: What are Europe’s main assets in supercomputing?

Dr. Kalbe: Two things come immediately to mind. First, Europe has become very good at sharing competencies and expertise across our nations. We’ve established good practices for doing this, not just in HPC but in many areas of common interest.  In HPC, Europe has world-class strengths in applications and other software. That’s a definite advantage we intend to keep by investing in user-driven applications and pushing them toward exascale performance. We also have centers-of-excellence to ensure users always have access to the best-available applications. Another European strength is our strong focus on co-design, on ensuring that HPC hardware and software are developed together with applications performance in mind. The EU has supported important activities in this area in recent years.

HPCwire: What are the biggest challenges Europe faces in HPC?

Dr. Kalbe: One of our biggest challenges, as I mentioned earlier, is that we don’t produce some key technologies for HPC, such as microprocessors.  Globally today, there’s not much choice in microprocessors. Being able to produce them ourselves is critical for Europe’s strategic independence in HPC and for advancing applications. We can’t co-design European HPC systems without European processors. Another challenge is the differences in HPC skills levels in Europe. Some countries have very advanced HPC communities and tools, but other regions aren’t at the same level. So, we need to make progress in ensuring that users throughout the EU have comparable levels of access and service and skills. SMEs are especially important to support in this way, as are some science communities in some countries.

HPCwire: When will the EuroHPC Joint Undertaking be operational?

Dr. Kalbe:The Joint Undertaking is already operational. By the end of next year, plans are for the JU to be moving forward at cruising speed.

HPCwire: What are the main success criteria for the Joint Undertaking?

Dr. Kalbe: The primary criterion is making globally competitive supercomputers and services available on schedule. A second success factor is managing to satisfy user needs and have enough capacity to run their applications at desired large scales. A third factor is building the European supply chain, not just in HPC but also for mass markets like automated driving systems, the IoT and commercial enterprise data centers.

HPCwire: What is the plan for EuroHPC supercomputer procurements?

Dr. Kalbe: We first plan to launch the procurements of at least two pre-exascale machines for deployment in 2020. The EuroHPC Joint Undertaking will own those machines but we will delegate their operation to existing HPC centers in Europe. Early next year we will select the HPC centers that will host these pre-exascale computers. Each one might be operated by a single center or a consortium. Once the hosting entities have been selected, we’ll proceed with the procurements and acceptance tests and the host sites will provide the access.

HPCwire: How will the funding work?

Dr. Kalbe: The pre-exascale machines will have joint funding, with up to 50% of the costs and an equal percent of the access time for European users, and the remainder paid for by the hosting states for their own usage. For the pre-exascale systems, the EU will pay for up to half of both the acquisition and operating costs. At the same time the pre-exascale procurements are happening, we’ll also call for proposals for at least two petascale machines to raise HPC capabilities in  certain countries. These machines will be hosted in HPC centers in those countries, and the JU will pay for up to 35% of the costs and use the same percentage of access time on the machines. Finally, we have already started working on next the next financial framework to start in 2021, which will cover the exascale machine procurements and development work for post-exascale technologies.

HPCwire: How will user requirements be built into these procurements?

Dr. Kalbe: We have advisory groups. Two of them were established under the JU that will reach out to user communities for recommendations on which user requirements should be addressed and the best way to do that for the machines we intend to buy. Those requirements will then be baked into calls for these machines. We are trying to avoid duplication where possible, so we want the pre-exascale machines to serve different applications but also complement the workloads being run on national supercomputers in Europe. The goal is to connect the EuroHPC pre-exascale and exascale machines with European national machines that will be available through PRACE, in such a way that there’s minimal replication of what the PRACE machines are very strong at handling. Of course, this will depend on everyone’s willingness to coordinate, but that’s the goal.

HPCwire: How important for EuroHPC is support for European industries?

Dr. Kalbe: This is a delicate question because EuroHPC is funding a public infrastructure with public money and we need to avoid stepping into the market place. We have the potential to offer up to 20% of the machines’ capacities to commercial firms for research they would not need to publish. If we do that, we would need to charge market prices. To gain free access, of course, commercial firms would need to publish the findings of their research. We might see the use of both these models by some automakers and other manufacturers. But we can’t provide more than 20% of the cycles and we can’t undercut market pricing.

HPCwire: Which European industries stand to benefit most from EuroHPC?

Dr. Kalbe: Certainly car manufacturers and other large engineering companies, along with oil and gas firms, logistics companies, pharmaceutical, aerospace and IT companies. It’s almost better to ask if there are any industries that would not be interested.

HPCwire: What about SMEs? Will EuroHPC help them, too?

Dr. Kalbe: Yes, there will be a continuation of what we’ve already started. First, we’ll have a dedicated action where we reach out to SMEs and have them identify business cases that would benefit from HPC use. SMEs are very numerous and usually very local, so we’ll launch more competence centers so that each EuroHPC participating state has at least one. That way, if an SME requests digital services, the company can work with the same center in their country from start to finish. We’ve had workshops on how to support SMEs. They are not so interested in hiring IT experts but they want someone who understands their business and can help them make the business case. We’ll continue the successful Fortissimo initiative and expand it beyond its current focus on manufacturing.  And we’ll establish more competence centers.

HPCwire: What is the relationship between EuroHPC and PRACE?

Dr. Kalbe: We see PRACE as an essential partner. We are counting on PRACE for two main reasons. First, PRACE provides access to Tier-0 machines and support services for European researchers, with support from the EU. We want to keep that. Second, JU machines will be operated by their hosting entities, but PRACE does scheduling very well and our intent is for PRACE to do this for the European machines, too.  Of course, they need to keep separation between the national and European resources.

HPCwire: How will EuroHPC balance support for creating a strong European HPC supply chain with the need of European scientists and engineers to have access to the best HPC systems, no matter where in the world they come from? How open will the EuroHPC procurements be to suppliers based in other parts of the world?

Dr. Kalbe: This has been the number one question that’s been raised over and over again by some European states and other interested parties. Our highest priority is to provide access for users to world-class machines. We won’t acquire these machines just to channel money to European vendors.  Our hope is that by 2022, European technology will be mature enough to be globally competitive. But legally, our calls need to be open to any vendor and we can’t limit them to vendors from Europe. Non-European vendors can apply at the same level as European ones.

HPCwire: I understand that procurements for experimental machines could legally be limited to certain vendors.

Dr. Kalbe: Our current plans are for all the procurements to be for production machines, not experimental machines. We want to procure exascale machines in 2022. We have a number of initiatives in Europe to develop missing technologies and we’ll have to wait and see whether they are competitive in time for the procurements. We’ll have an extreme-scale demonstrator machine to see whether the design for the European processor is stable, for example. But whether it is or not, the pre-exascale and exascale acquisitions will go on and they will be open calls for production machines. European competitors will need to compete strictly on merit, although ideally we would like companies based in the EU to be able to develop the machines. If a major vendor not based in Europe wants to work with European suppliers to finish development of the machines, that would be a very good approach from our perspective. For the next two years, it will be a challenge to have all the necessary European components in place, but we’re hoping that by 2022, we’ll be able to launch a call for at least one European machine.

HPCwire: There is an HPC skills gap throughout the world. How will Europe address the skills gap?

Dr. Kalbe: This is a major challenge that we are addressing with a number of instruments. For the next financial framework, we will increase investment in digital skills, not just in HPC. Linked to this will be the establishment of the competence centers. They will have as part of their missions to provide training and competencies so we have a stronger, larger skills base in Europe and in all the member states. More indirectly, by increasing the available capacity of the supercomputers, more users can be satisfied within the EU instead of going outside, so their skills stay in Europe.

HPCwire: The European Processor Initiative has established ambitious goals to achieve by 2025, including the development of an HPC processor, an embedded processor for automated vehicles and other AI applications, and an accelerator based on the RISC-V standard. How will the EuroHPC Joint Undertaking be affected if EPI doesn’t deliver results on time?

Dr. Kalbe: We’re not putting all our eggs in one basket. The EPI is major initiative not just for HPC but also for the mass market and auto companies that are part of this consortium. We’ll explore alternative approaches so we’ll have alternatives in case EPI takes longer. We’ll see in 2022, when we launch calls for the exascale machines, whether we will have mature European processors or need to do an open call for processors.

HPCwire: Finally, the EU plans to invest one billion euros over ten years in the quantum computing flagship program that is now under way. Will EuroHPC be involved in this program or in any aspect of quantum computing?

Dr. Kalbe: We see more and more convergence between HPC and quantum R&D. We have distinct programs for them today, but by 2022 we plan to link them together formally. We are hoping for a hybrid HPC-quantum machine after 2025. We’ve put a plan in place to make experimental quantum computers available to users for developing codes, to see how hybrid machines might be built as precursors to pure quantum machines.

HPCwire: What’s the expectation for the state of EU in the HPC sector by 2025?

Dr. Kalbe: In 2025, we should have a fully interconnected infrastructure across Europe in HPC, combining national and European machines. The fully meshed network will be available to any user in Europe, with a single point of entry where the user sends a request and it goes to the most appropriate machine and this process is completely transparent to the user. The user won’t need to care where the application runs, whether it’s on a national or European machine. The second expectation is that we manage to keep European science and industry competitive because they have had access to world-class HPC resources.

Subscribe to HPCwire's Weekly Update!

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

A Beginner’s Guide to the ASC19 Finals

April 22, 2019

Three thousand watts. That's how much power the competitors in the 2019 ASC Student Supercomputer Challenge here in Dalian, China, have to work with. Everybody would like more juice to run compute-intensive HPC simulatio Read more…

By Alex Woodie

Is Data Science the Fourth Pillar of the Scientific Method?

April 18, 2019

Nvidia CEO Jensen Huang revived a decade-old debate last month when he said that modern data science (AI plus HPC) has become the fourth pillar of the scientific method. While some disagree with the notion that statistic Read more…

By Alex Woodie

At ASF 2019: The Virtuous Circle of Big Data, AI and HPC

April 18, 2019

We've entered a new phase in IT -- in the world, really -- where the combination of big data, artificial intelligence, and high performance computing is pushing the bounds of what's possible in business and science, in w Read more…

By Alex Woodie with Doug Black and Tiffany Trader

HPE Extreme Performance Solutions

HPE and Intel® Omni-Path Architecture: How to Power a Cloud

Learn how HPE and Intel® Omni-Path Architecture provide critical infrastructure for leading Nordic HPC provider’s HPCFLOW cloud service.

powercloud_blog.jpgFor decades, HPE has been at the forefront of high-performance computing, and we’ve powered some of the fastest and most robust supercomputers in the world. Read more…

IBM Accelerated Insights

Bridging HPC and Cloud Native Development with Kubernetes

The HPC community has historically developed its own specialized software stack including schedulers, filesystems, developer tools, container technologies tuned for performance and large-scale on-premises deployments. Read more…

Google Open Sources TensorFlow Version of MorphNet DL Tool

April 18, 2019

Designing optimum deep neural networks remains a non-trivial exercise. “Given the large search space of possible architectures, designing a network from scratch for your specific application can be prohibitively expens Read more…

By John Russell

A Beginner’s Guide to the ASC19 Finals

April 22, 2019

Three thousand watts. That's how much power the competitors in the 2019 ASC Student Supercomputer Challenge here in Dalian, China, have to work with. Everybody Read more…

By Alex Woodie

At ASF 2019: The Virtuous Circle of Big Data, AI and HPC

April 18, 2019

We've entered a new phase in IT -- in the world, really -- where the combination of big data, artificial intelligence, and high performance computing is pushing Read more…

By Alex Woodie with Doug Black and Tiffany Trader

Interview with 2019 Person to Watch Michela Taufer

April 18, 2019

Today, as part of our ongoing HPCwire People to Watch focus series, we are highlighting our interview with 2019 Person to Watch Michela Taufer. Michela -- the Read more…

By HPCwire Editorial Team

Intel Gold U-Series SKUs Reveal Single Socket Intentions

April 18, 2019

Intel plans to jump into the single socket market with a portion of its just announced Cascade Lake microprocessor line according to one media report. This isn Read more…

By John Russell

BSC Researchers Shrink Floating Point Formats to Accelerate Deep Neural Network Training

April 15, 2019

Sometimes calculating solutions as precisely as a computer can wastes more CPU resources than is necessary. A case in point is with deep learning. In early stag Read more…

By Ken Strandberg

Intel Extends FPGA Ecosystem with 10nm Agilex

April 11, 2019

The insatiable appetite for higher throughput and lower latency – particularly where edge analytics and AI, network functions, or for a range of datacenter ac Read more…

By Doug Black

Nvidia Doubles Down on Medical AI

April 9, 2019

Nvidia is collaborating with medical groups to push GPU-powered AI tools into clinical settings, including radiology and drug discovery. The GPU leader said Monday it will collaborate with the American College of Radiology (ACR) to provide clinicians with its Clara AI tool kit. The partnership would allow radiologists to leverage AI techniques for diagnostic imaging using their own clinical data. Read more…

By George Leopold

Digging into MLPerf Benchmark Suite to Inform AI Infrastructure Decisions

April 9, 2019

With machine learning and deep learning storming into the datacenter, the new challenge is optimizing infrastructure choices to support diverse ML and DL workfl Read more…

By John Russell

The Case Against ‘The Case Against Quantum Computing’

January 9, 2019

It’s not easy to be a physicist. Richard Feynman (basically the Jimi Hendrix of physicists) once said: “The first principle is that you must not fool yourse Read more…

By Ben Criger

Why Nvidia Bought Mellanox: ‘Future Datacenters Will Be…Like High Performance Computers’

March 14, 2019

“Future datacenters of all kinds will be built like high performance computers,” said Nvidia CEO Jensen Huang during a phone briefing on Monday after Nvidia revealed scooping up the high performance networking company Mellanox for $6.9 billion. Read more…

By Tiffany Trader

ClusterVision in Bankruptcy, Fate Uncertain

February 13, 2019

ClusterVision, European HPC specialists that have built and installed over 20 Top500-ranked systems in their nearly 17-year history, appear to be in the midst o Read more…

By Tiffany Trader

Intel Reportedly in $6B Bid for Mellanox

January 30, 2019

The latest rumors and reports around an acquisition of Mellanox focus on Intel, which has reportedly offered a $6 billion bid for the high performance interconn Read more…

By Doug Black

It’s Official: Aurora on Track to Be First US Exascale Computer in 2021

March 18, 2019

The U.S. Department of Energy along with Intel and Cray confirmed today that an Intel/Cray supercomputer, "Aurora," capable of sustained performance of one exaf Read more…

By Tiffany Trader

Looking for Light Reading? NSF-backed ‘Comic Books’ Tackle Quantum Computing

January 28, 2019

Still baffled by quantum computing? How about turning to comic books (graphic novels for the well-read among you) for some clarity and a little humor on QC. The Read more…

By John Russell

IBM Quantum Update: Q System One Launch, New Collaborators, and QC Center Plans

January 10, 2019

IBM made three significant quantum computing announcements at CES this week. One was introduction of IBM Q System One; it’s really the integration of IBM’s Read more…

By John Russell

Deep500: ETH Researchers Introduce New Deep Learning Benchmark for HPC

February 5, 2019

ETH researchers have developed a new deep learning benchmarking environment – Deep500 – they say is “the first distributed and reproducible benchmarking s Read more…

By John Russell

Leading Solution Providers

SC 18 Virtual Booth Video Tour

Advania @ SC18 AMD @ SC18
ASRock Rack @ SC18
DDN Storage @ SC18
HPE @ SC18
IBM @ SC18
Lenovo @ SC18 Mellanox Technologies @ SC18
One Stop Systems @ SC18
Oracle @ SC18 Panasas @ SC18
Supermicro @ SC18 SUSE @ SC18 TYAN @ SC18
Verne Global @ SC18

IBM Bets $2B Seeking 1000X AI Hardware Performance Boost

February 7, 2019

For now, AI systems are mostly machine learning-based and “narrow” – powerful as they are by today's standards, they're limited to performing a few, narro Read more…

By Doug Black

The Deep500 – Researchers Tackle an HPC Benchmark for Deep Learning

January 7, 2019

How do you know if an HPC system, particularly a larger-scale system, is well-suited for deep learning workloads? Today, that’s not an easy question to answer Read more…

By John Russell

Arm Unveils Neoverse N1 Platform with up to 128-Cores

February 20, 2019

Following on its Neoverse roadmap announcement last October, Arm today revealed its next-gen Neoverse microarchitecture with compute and throughput-optimized si Read more…

By Tiffany Trader

Intel Launches Cascade Lake Xeons with Up to 56 Cores

April 2, 2019

At Intel's Data-Centric Innovation Day in San Francisco (April 2), the company unveiled its second-generation Xeon Scalable (Cascade Lake) family and debuted it Read more…

By Tiffany Trader

France to Deploy AI-Focused Supercomputer: Jean Zay

January 22, 2019

HPE announced today that it won the contract to build a supercomputer that will drive France’s AI and HPC efforts. The computer will be part of GENCI, the Fre Read more…

By Tiffany Trader

Oil and Gas Supercloud Clears Out Remaining Knights Landing Inventory: All 38,000 Wafers

March 13, 2019

The McCloud HPC service being built by Australia’s DownUnder GeoSolutions (DUG) outside Houston is set to become the largest oil and gas cloud in the world th Read more…

By Tiffany Trader

Intel Extends FPGA Ecosystem with 10nm Agilex

April 11, 2019

The insatiable appetite for higher throughput and lower latency – particularly where edge analytics and AI, network functions, or for a range of datacenter ac Read more…

By Doug Black

UC Berkeley Paper Heralds Rise of Serverless Computing in the Cloud – Do You Agree?

February 13, 2019

Almost exactly ten years to the day from publishing of their widely-read, seminal paper on cloud computing, UC Berkeley researchers have issued another ambitious examination of cloud computing - Cloud Programming Simplified: A Berkeley View on Serverless Computing. The new work heralds the rise of ‘serverless computing’ as the next dominant phase of cloud computing. Read more…

By John Russell

  • arrow
  • Click Here for More Headlines
  • arrow
Do NOT follow this link or you will be banned from the site!
Share This