In November we ran a BoF at Supercomputing 18 (SC18) entitled “HPC Meets Real-Time Data: Interactive Supercomputing for Urgent Decision Making”. This was really the first piece of dissemination activity of the VESTEC project and, as there are no other current dissemination activities, such as BoFs or workshops, in the HPC community around urgent decision making, I was somewhat nervous about how many attendees we were likely to get and how engaged they would be! Out master plan, or what we were looking to do, really goes a bit beyond the VESTEC project itself, and was instead looking at taking the first step in building a global community based around using HPC for urgent decision making, with VESTEC being a part of this.
Even as early as proposing the BoF, it felt fairly natural to split the one hour time slot up into three concurrent group activities. The idea being that each group would hold a discussion focussing on a distinct area of interest and then we would bring together all participants at the end for a short wash-up. Why three topics you might ask? Well, these were the applications that can benefit from using HPC for urgent decision making, challenges arising from fusing HPC with real time data for urgent decision making, and visualisation of the outputs such that urgent decision makers can make the correct decision first time, every time. We felt that, from our perspective, these are three major pillars of using HPC for urgent decision making (although, as we found out during the BoF, there are others!)
I’ve got to say, even though the BoF was at lunchtime, and-so we were competing against people’s empty stomachs, it went really well, with over sixty participants at some points. Each discussion group had a good number of engaged people and the thirty minutes we had allocated to this part of the BoF went really quickly! From my position it was quite clear that we could have easily filled up another hour-long slot exploring the challenges that were identified.
Going into this BoF we know that HPC for urgent decision making could be a hot topic, with a real world impact, but given the lack of existing community activities around this area I was surprised at just how many people were interested in the topic and that already many people have their own ideas, use-cases, and interests that can contribute to this area. I think this really illustrates that the VESTEC project is looking to tackle important challenges and people already understand that, if we get this right, there could be significant real-world benefit.
So what about my overarching ambition to take the first step in building a global community in urgent decision making? Well we made numerous contacts through the BoF and have already identified a number of other projects worldwide that we believe are complementary in this area. It really makes sense to further build these connections, and as a group we are looking for opportunities to take the next step in building the HPC for urgent decision making community.
We have written a detailed summary on the discussions held and points raised during the BoF, read it here.
Nick Brown, EPCC