Georgia Institute of Technology, United States of America
Depending on how you define it, Moore's Law has ended, is about to end, or will virtually never end. Regardless of the definition, it's clear that we are in the midst of a period of significant change in computer architectures. Riding the semiconductor innovation curve is a distant memory. But what comes next? Yes, the hype over the non-von Neumann computation approaches of neuromorphic and quantum computing has reached new heights. But are there other approaches on the horizon? If so, what are they? How much will they disrupt current algorithms and software infrastructures?
In 2013, along with Elie Track, I co-founded the IEEE Rebooting Computing Initiative (RCI). Since that time, the initiative has held invitational blue sky sessions, open panels and, since 2016, an open conference series (the International Conference on Rebooting Computing). Throughout this journey, we've heard discussion covering the spectrum from doom-and-gloom to promising predictions for the future of HPC. This talk will review the road ahead, as seen from the IEEE RCI perspective, and suggest some promising technologies on which to keep a close eye.