HPCwire presents our interview with Eric Monchalin, chair of European Processor Initiative & VP, head of machine intelligence, Atos, and an HPCwire 2022 Person to Watch. Monchalin recaps the EPI program strategy, explains the motivations for tech sovereignty, and gives a view to current and future computing trends. This interview was conducted by email earlier this year.
Congratulations on being named a 2022 HPCwire Person to Watch! The European Processor Initiative seems more pressing than ever, with the pandemic revealing the fragility of many international supply chains that are crucial for modern computing and manufacturing. How do you view EPI’s role in such a volatile landscape?
Thank you for this honor. I am delighted to be recognized as a 2022 HPCwire Person to Watch. I’m equally delighted to have been elected chair of the European Processor Initiative and to actively participate, also as a European citizen, to pave the way toward Europe’s technological sovereignty.
The pandemic has underlined how countries and economical spaces, like the European Union, are depending on other geographical areas for their key procurements. Even with highly sophisticated logistics chains delivering just-in-time parts and goods, it has been demonstrated that any severe, unplanned event can lead to serious shortages strongly impacting economies or business systems.
The dramatic consequences of these shortages, plus among other things like the net zero target or the relentless increase of the cost of energy, is raising the importance of localizing sustainable, efficient and resilient manufacturing capabilities. It’s also pointing out the value of smart manufacturing, also named Industry 4.0, to achieve such objectives.
It is a big challenge in itself to bridge the gap between the European strengths and challenges to make Europe a leader in the Industry 4.0, and thus to bring a European localized industry to the forefront. However, much more is possible when a concerted academic, research and industrial effort is developed to break out of the technological locks. It is exactly the objective of the European Processor Initiative, with 30 partners from 11 European countries cooperating, to strengthen the competitiveness and leadership of European industry and science, designing European general purpose and accelerator microprocessor technologies with extreme performance and power ratio, as well as tackling important segments of broad and emerging HPC and Big Data markets.
The European Processor Initiative is ultimately targeting the coming European Exascale systems, as well as the computing continuum, as the foundation of Industry 4.0 to pave the road of a leading, resilient and sovereign European industry.
Initiatives like EuroHPC and the European Processor Initiative are working to advance European sovereignty in supercomputing. How do you view the balance between (and future of) global, continental and national interests in European supercomputing?
Europe has a long history in mathematics and science with strong academic institutions. It explains the European leadership on the application field. However, Europe was less focused on the hardware technologies with nowadays being dependent of non-European players for the design and manufacturing of many key components like memories or high-end processors. The good news is that European academic and industrial actors are leaders in other types of components, providing a fertile breeding ground for the European Chips Act that has been announced on 8th February.
The European Processor Initiative is in some ways a forerunner of the European Chips Act. EPI contributes to the EU digital agenda and strengthens Europe’s digital autonomy, and it serves the objective of EuroHPC, which is building a global, leading European HPC ecosystem. The European Processor Initiative’s impacts are even twofold. First, it contributes to the development of European supercomputing technologies that can compete on the global HPC market. Second, it is a key ingredient of the European Union to equip itself with a world-class supercomputing infrastructure.
The EuroHPC roadmap (2019-2027) includes the integration of quantum and HPC infrastructures. Does Europe see this as a leadership opportunity, in what ways?
If Europe has to improve its investments in chips for supercomputing infrastructure to strengthen its leadership and resilience, the situation is totally different for quantum computing. All worldwide players are almost starting the race on an equal footing, with equal chances of being first in class.
As everywhere, the quantum field is therefore strongly active in Europe with large European investments in the Quantum Flagship launched in 2018 and ambitious national programs and numerous promising academic, research, startup and industrial players. Atos is by example collaborating with the startup Pasqal to develop a quantum accelerator for supercomputers. Atos is also making available to the market its Quantum Learning Machine, the highest performing quantum simulator in the world.
Nevertheless, it is still a long path in front of us to see universal quantum computers that are reliable enough for around-the-clock science or industrial production. In a first step, it’s a promising technology as an accelerator to boost HPC applications in areas like optimization problems, chemistry or machine learning.
Consequently, quantum computing has full importance in the EuroHPC roadmap to equip the European Union “with a world-class federated, secure and hyper-connected supercomputing and quantum computing service and data infrastructure” as stated by Anders Jensen in his HPCwire People to Watch recognition from last year’s article.
Where do you see HPC headed? What trends – and in particular emerging trends – do you find most notable? Any areas you are concerned about, or identify as in need of more attention/investment?
The pace of technological change is faster than ever with a digital revolution under way for which the drivers are proliferating: HPC, Edge Computing, Internet of Things, AI, Industrial Data Platforms, Blockchain and so on. These drivers are enabling the transformation of manufacturing and services, reshaping entire sectors of the economy. Associated to strong academic institutions and industrial infrastructures, these drivers are pillars for the growth of ecosystems of smart machines that is a foundation for Smart Manufacturing or Industry 4.0 – the underway industrial revolution. These swarm computing capabilities are also enabled by the datacenters, Edge Computing and IoT that become increasingly connected, emerging as a computing continuum.
A complete and efficient industrial product or process is also being accompanied by a virtual representation: a digital twin. These digital solutions shorten the long iterative design and manufacturing cycles, lessening the time-to-market by removing idle time when waiting for information. Digital Twins are a great opportunity in utilizing the power of models in combination with massive computing capacities, the increasing availability of rich data sets, modern data analytics and AI techniques.
In such a context, supercomputing is broader than ever, embracing classical modeling and simulation that are increasingly boosted by Machine Learning models. Data driven models are also injected in numerical simulations to represent complex systems, and massive outputs of large simulations require nowadays AI capabilities to analyze the results. In addition, real-time supercomputing is needed by more and more fields like Industry 4.0 and digital twins. Lastly a green computing continuum is pushing supercomputing at the edge.
These massive and fast progresses are reinforcing the key role of supercomputing for all the societal and business challenges like climate change, decarbonation, sustainability, smart everything (cities, mobility, agriculture, medicine, manufacturing etc.), energy challenges and so on. It’s not only a must for any continent or economical space or nation to tackle these numerous challenges by their own to strengthen their leadership and autonomy, but also in cooperation with others.
Lastly, Atos is actively participating to these challenges with most recently the announcement of the BullSequana XH3000, a hybrid computing platform with uparalleled flexibility, performance and energy-efficiency to meet the Exascale challenges above and beyond.
Outside of the professional sphere, what can you tell us about yourself – family stories, unique hobbies, favorite places, etc.? Is there anything about you your colleagues might be surprised to learn?
I’m fortunate to live in Grenoble within the Alpes. The closest starting points for local summits, ski stations or windy lakes are only 30 minutes away from home. To live here was a big opportunity during the pandemic lockdowns to have the ability to relax, hiking from home on nearby mountainsides. In such a great area you must love outdoor activities with your family. We enjoy hiking and climbing. When we notice the snow on the peaks, it’s time to put on either the skis, the snowboard or, for most of the time, the snowshoes.
I also have secret to admit. The water is always too cold for me. No chance to see me swimming somewhere. But I do love sailing. There is no better shot of pleasure and adrenaline than when you’re sailing fast enough that the boat is planning on top of the water under a large spinnaker in strong wind conditions. Doing this alongside one of my adult children on the family sport boat, being splashed by the hull just above the water surface, is a thrill I can’t get enough of.
The weather can also be inclement. In this case, it’s time for creative activities at home like developing a robot, practicing bookbinding, or spending some time in my photo lab, developing an old fashion black and white film and making large format prints from my best shots.
Monchalin is one of 12 HPCwire People to Watch for 2022. You can read the interviews with the other honorees at this link.