NCSA Director Bill Gropp on SC23, Future Path of HPC

Editor’s note: This is part of a series of virtual visits with NCSA thought leaders on current topics impacting the field of high-performance computing.

SC23 and the Future Path of HPC
By Bill Gropp, NCSA Director

The transition to a new year is the perfect time for reflection, whether it be personally or professionally. The opportunity to attend SC in November each fall serves as a convenient bridge from the year that was in high-performance computing to turning possibilities into realities in the future.

Much was accomplished at the National Center for Supercomputing Applications and throughout the greater advanced research computing community in 2023. I was especially proud of the recognition our teams received from the HPCwire Readers’ and Editors’ Choice Awards because it’s an example of the great partnerships NCSA has built and the impact of the initiatives we’ve undertaken. It was clear NCSA has continued to be a leader in forward-looking architectures and applications, and conversations with other potential collaborators at SC will most certainly lead to additional opportunities for NCSA to shine.

The people and partnerships found within the HPC community were a prevalent theme throughout SC23 but was perhaps most apparent in the keynote presentation by Dr. Hakeem Oluseyi, an astrophysicist and former Space Science Education Lead for NASA. I found his speech so inspiring. The opportunity we have to find people who are as passionate about what we do – it’s great not only to find those people but to help them succeed. It was also a great reminder that people can make some wrong turns – I think he was really great and forthright about that – but that they can still change and not just survive but excel. 

It was a different kind of talk than conferences in the past, but it fit in well with the “I Am HPC” theme of SC23. I really believe in and support that vision because what’s wonderful about this community are the people. What Dr. Oluseyi is doing is challenging us to be even better at being inclusive, being open and bringing new people into the community.

In some sense, selecting somebody outside of HPC who could touch us was part of that message, I think. It wasn’t an inside-computing person but rather a person outside of HPC who sees that what we do is important and contributes to advancing science and society.

The exhibit floor at SC23 gave a perfect snapshot of where HPC is now and where it is going. Highly notable were the number of booths – both from industry and research organizations – that discussed artificial intelligence, GPUs and specialized architectures; better I/O; HPC in the cloud; and, especially, cooling technologies. SC itself set records for the number of exhibitors and attendees, showing the strength of the field of high-performance computing.

Bill Gropp, NCSA Director

I was impressed by the strong showing from commercial cloud vendors. They’ve attended the conference before, but this time there was a large presence for HPC in the cloud by all of the major vendors. It was a good reminder that they provide other capabilities that complement traditional HPC approaches.

So what’s next? What will be the big trend that stands out at SC24? Will AI continue to speed ahead? What will quantum computing look like six, nine, 12 months from now?

I tend to shy away from making predictions, but I do think we’ll continue to see more use of machine learning and AI, as well as a growing understanding of the limitations of machine learning. We’ll continue to see a big gap between the peak and achieved performance of increasingly complex systems, which is an opportunity for human-machine partnerships in programming approaches. And, as always, I’ll be looking for new computer architectures to break out at SC24 in Atlanta.