Round up | On the trail to Exascale and Scalable AI

As a part of the STFC Hartree Centre’s goal to help UK businesses and organisations explore cutting-edge technologies, we recently hosted an exascale workshop at Liverpool Town Hall.     

From clean energy and decarbonising the economy to drug and material discovery, exascale computing has the potential to transform how we tackle grand challenges with its ability to process larger quantities of data faster than ever before, helping to accelerate innovation.

For the second year in a row our On the trail to Exascale workshop brought together representatives from industry, academia and the scientific community to discuss how to help the UK move forward in the exascale landscape. We created a space for industry to talk and collaborate with visitors from the Department for Science, Innovation and Technology (DSIT), UK supercomputing centres, and representatives from the US National Labs, Lawrence Livermore National Laboratory (LLNL) and Oak Ridge National Laboratory (ORNL). It was exciting to see individuals from a range of sectors collaborate and open discussions on how to enhance our quality of life with emerging HPC technologies, and how they can help businesses tackle global challenges. 

“One of the key points to come out of our discussions is the ability of large-scale compute and exascale computing to address the grand challenges that our society faces. Some of these grand challenges are being able to fully simulate clean energy devices and predict climate patterns to a very fine detail. What stood out for me is how, in addressing such complex problems, collaboration is the key to fully harness the potential of different communities and address a range of challenges. As a part of this, we are collaborating with the US national labs to bring together experts from different domains to see how we can move towards harnessing the power of exascale computing.” 

Ubaid Ali Qadri, Team Leader for Multi-Fidelity Design and Twinning 

On the left Erik Draeger, Lawrence Livermore National Lab and on the right Vassil Alexandrov.

The workshop kicked off with Vassil Alexandrov, our Chief Science Officer and Erik Draeger, Lawrence Livermore National Lab sharing how the journey to exascale computing has developed over the last year and what this means for industry. Exascale computing technologies can help organisations model more detailed simulations faster than ever before, helping to accelerate innovation.   

“We are working with colleagues from the US National Labs, focusing on scalable AI to exploit opportunities with high societal impact. We are also partnering with several UK exascale computing projects through the ExCALIBUR programme, like the MetOffice to optimise their climate models.” 

Vassil Alexandrov, Chief Science Officer 

On the right Mark Wilkinson, DiRAC and on the left Valerie Farr, STFC.

With the growing need for exascale computing and AI across industry and the research community STFC is developing strategies to support innovation and help grow the exascale community in the UK. Over the course of the two days we heard from Mark Wilkinson, DiRAC and Valerie Farr, STFC who spoke about their respective exascale and AI strategies and how they are identifying key community needs within their working groups. Mark Wilkinson, explained how he has worked with our Chief Science Officer, Vassil Alexandrov, to host STFC exascale “town hall meetings” to help develop our exascale requirements and make the subject accessible for everyone in science and industry.

“The key thing about exascale is that it needs a clear ecosystem, which supports organisations at all levels, and STFC is investing in people and making sure that they can get the most out of it. Exascale computing is about scaling up your work, whether that is increasing your datasets by a 1000 times or by optimising your processing speed by 1000 times, it is for everyone!”

Mark Wilkinson, DiRAC

Valerie Farr, who is part of the working group developing STFC’s AI research focus, emphasised the importance of people-centred technology.  

“We are at the phase where we really can understand the opportunities and possibilities of integrating AI across our research centres. People are absolutely critical to the development and sustainability of AI.”

Valerie Farr, STFC Strategy, Planning and Communications 

With our growing need for more powerful computers to tackle grand challenges, it is important to design them sustainably to reduce their environmental impact. Sadaf R. Alam, University of Bristol’s Director of Advanced Computing Strategy, detailed how they focused the design specifications for their new AI Research Resource (AIRR) computer on sustainability.

“As we build more powerful machines, we need to do so sustainably and make them accessible to all users.”

Sadaf R. Alam, University of Bristol

Over the second day we heard from some of our experts about how we are helping organisations to explore and adopt emerging technologies like scalable AI helping industry and the public sector become exascale ready. 

“One of the ways we have supported companies to use AI is material discovery, such as simulating the best materials for carbon capture. We have also helped organisations across different sectors to speed up and optimise their design process, from improving simulations of values for hydrogen transport to modelling a fusion energy device.”

Sue Thorne, Team Leader and Computational Scientist

The Department of Science, Innovation and Technology (DSIT) also featured at the event, sharing the government’s plans to drive opportunity for AI across the UK and how they are looking to build our AI computing capability and skills. The workshop was rounded off with a lively panel discussion about the future adoption of exascale and AI technologies across industries, and our next steps. 

“Moving forwards high-precision HPC modelling and simulation will be critical to developing scientific data sets for next phase of AI. This will help us to create digital twins for manufacturing and precision medicine or designing energy storage and carbon capture.”

Brian C. Van Essen, Lawrence Livermore National Laboratory

We are proud to be helping industry, the public sector and the research community to explore and adopt cutting-edge technologies like scalable AI to help them become exascale-ready. To keep in touch as we build an exascale ecosystem in the UK to help tackle grand challenges develop effective solutions and accelerate innovation, you can follow our social media channels or subscribe to our newsletter.