• Subscribe

What's next for supercomputing in Europe?

Last week, Science Node traveled to Poznań, Poland to find out what our friends in the European HPC community are up to. EuroHPC Summit Week/PRACEdays19 brings together researchers, computer scientists, and industry partners to share successes and prepare for the future of high-performance computing (HPC).

HPC on the horizon. EuroHPCSummit Week/PRACEdays19 looks ahead to the future of scientific computing in Europe--from exascale to industry partnerships.

This year, much discussion was given to the changes coming to the HPC landscape. As we’ve written about elsewhere, exascale computers are very much on the horizon. But at the same time, the end of Moore’s Law means an end to rapid performance acceleration—at least in its familiar forms.

Taken together, these two factors mean this is a time of both uncertainty and possibility for HPC. On the whole, the researchers we spoke with were eager to confront the new era in computing.

As Núria López, chair of the PRACE Scientific Steering Committee, described it, “We all have these projects we keep in our pocket because we don’t yet have the resources to investigate them.” But she awaits the day when the promised new capabilities finally do arrive and all of those reserved projects come out of the “someday” and into the now.

From AI to industry 

<strong>David Keyes</strong> discusses how the evolution of machine learning and high-end data analytics are influencing HPC as we move into the exascale era. Courtesy PRACE. David Keyes, Director of Extreme Computing Research Center at King Abdullah University of Science and Technology presented about the impacts of big data and machine learning on extreme-scale computing.

As Keyes pointed out, a big data paper won last year’s Gordon Bell Prize for the first time, and half of the finalists were working with big data. He suggested that while science and engineering fields may be minority users of big data, they still have the capacity to become leaders in the big data community by harnessing HPC for new uses and by continuing to be pathfinders for new applications. 

Ultimately, Keyes believes that the convergence of simulation and big data with exascale computation will give humanity the predictive tools necessary to overcome great natural and technical challenges.

<strong>On the edge.</strong> Rosa Badia presents use cases for smart cities and other projects incorporating data from edge sensors. Courtesy PRACE.Rosa Badia of Barcelona Supercomputing Center (BSC) presented three use cases combining supercomputing and big data: an indoor navigation system at Cagliaria airport in Italy; smart traffic management in Modena; and autonomous positioning, driving assistance, and predictive maintenance for a tramway network in Florence. All three projects have numerous sensors that collect data from the edge and present new challenges to infrastructure, data storage, and workflow orchestration.

Another strong theme of the conference was the importance of partnering with industry—and how best to do so. Brendan McGinty, Industry Director at the National Center for Supercomputing Applications (NCSA) in Illinois, shared insights from his center’s successful relationships with companies in aerospace, agriculture, health care, communications, finance, and insurance.

McGinty emphasized that for academic and research institutions to succeed with industrial partners sometimes requires a shift in mindset. He cautions that to industry, “deadlines are real” and must be met, and turnaround times are often very fast. However, he says, if industry partners receive value, they will keep coming back.

<strong>Oriol Lehmkul</strong> of BSC (r) receives the award for PRACE Best Industrial Paper from Lee Margetts (l) for his work simulating vortex-induced vibrations with the aim of improving wind turbines. Courtesy PRACE.Lest anyone think that working with industry doesn’t provide sufficient scope for scientific discovery, this year’s winner of the PRACE Best Industrial Presentation proves that false. BSC researcher Oriol Lehmkuhl’s paper on large-scale simulations of vortex-induced vibrations grew out of a SHAPE project to improve efficiency of wind farms. (SHAPE is a pan-European program supporting HPC adoption by small and medium-sized enterprises.) 

As judge Lee Margetts of the University of Manchester (UoM) noted when presenting the award, one reason the paper was chosen was because it was actually inspired by Lehmkuhl’s work with an industrial partner. “It demonstrates the importance of the interaction between science and industry for creation of new ideas and solution to problems that scientists haven’t thought about,” said Margetts.

Encouraging the next generation 

This year’s events also included an all-day session addressing concerns about equality and diversity in the HPC community. Rosa Badia presented about the successes of the ‘Somos investigadoras” program at BSC that aims to engage the attention and interest of young girls.

Somos investigadoras. “MareNostrum,” a song by the kids’ pop group Macedònia celebrates BSC’s supercomputer. The aim is to encourage children to study tech subjects and, in particular, to appeal to girls. Courtesy BSC CNS.

Lee Margetts shared examples of how UoM had achieved a degree of success, starting by changing policy but ultimately altering the culture of the institution.

A lively discussion and open conversation followed about actions and changes that could be taken to help all people feel more comfortable in the community. Several panelists emphasized the importance of promoting inclusivity as early as primary school to change attitudes about who does and doesn’t belong in STEM.

Additionally, a hands-on-workshop for students was organized by PRACE and held at host institution Poznań Supercomputing and Networking Center.

 

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2019 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.

Republish

We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.