Struggling to escape the collapse of Moore's Law

*Okay, there's a basic problem with the idea "rebooting computing," because Moore's Law didn't collapse for technological reasons. It didn't need a "reboot." Moore's Law is failing because faster and more powerful processors can't be made to pay.

*And it's not like you don't get cheap, increased computational power in a world without Moore's Law; you just rent computation from the cloud. A different paradigm.

*Moore's Law is ending now for many of the same reasons that the Space Age ended. It was technically possible to build larger, faster spacecraft and put more people on the Moon. But there's no economic model under capitalism for a manned moon colony or even a manned space station. So even if you build a completely radical form of increased launch capacity, like, say, a giant magic rubber-band slingshot in your backyard, you'll have a hard time rekindling the collapsed soufflé' of the golden-age of space-racing.

*That's not to say that space-racing's doomed forever, or can't be "rebooted." Space fervor would pop back pronto if there was a way to make it pay. Like, say, by scraping tons of rare metals off of asteroids. It could happen.

*There might also be some revival of the commercial motive to stuff a megaton of heavy-iron computational power into a small space. In which case, Moore's Law might rule the tech world again, and even become more rigorous than it was before.

*But that grim assessment of mine is not gonna daunt the engineering guys at the IEEE; no, telling them to give it up with the super-chips is just like telling science fiction writers to knock it off with writing faster-than-light galactic space-operas. So, instead of meekly despairing, they've come up with this cool "rebooting computing" effort to revive Moore's Law by coming up with completely alien paradigms for computation. It's baroque, outlandish, very sci-fi-ish, and as a mature contemporary adult I can't see much practical use for it, but personally, my enthusiasm for this project knows few bounds. I love these initiatives.

*They're so interesting and different. "Spintronic superconductive computing," man, I would have to be at death's door not to like that.

http://rebootingcomputing.ieee.org/archived-articles-and-videos/general/highlights-of-the-1st-ieee-international-conference-on-rebooting-computing-icrc-2016

ICRC 2016, sponsored by the Rebooting Computing Initiative (RCI) of the IEEE Future Directions Committee, was held 17-19 October, 2016, at the Hilton San Diego/Del Mar, in Del Mar, California. ICRC included 150 participants from the US, Europe, and Asia, with a goal of fostering continued improvements in future computer technology. The program for the conference is shown here. A recent overview of ICRC in IEEE Institute is show here.

Many of the talks are accessible from IEEE.tv, and many of the papers are included in the ICRC Proceedings on IEEE Xplore.

ICRC built on a series of 4 RCI Summits held from 2013-2015. The conference chair was Stan Williams of HPE, and the program chair was John Paul Strachan, also of HPE. A second ICRC is being planned for Oct. 2017 in Washington, DC. Please check back for further information on ICRC 2017.

The conference opened with a review of the history of computing by Dr. Robert Leland of Sandia, focusing on the early role of John Von Neumann in establishing traditional digital computer architectures in the 1950s, followed by Moore’s Law scaling of transistors since the 1960s. But both of these are reaching practical limitations: Delays and power are dominated by shuttling bits between logic and memory, and smaller CMOS transistors are no longer cheaper and more reliable.

In order to continue to improve performance, we need to “reboot” computer technology itself, at all levels. This will not require discarding what we have developed, but rather taking a second look at a variety of evolutionary and revolutionary approaches that offer ways around the current limitations. ICRC included papers on the following:

Neuromorphic or Brain-Inspired Computing
Approximate and Stochastic Computing
Adiabatic and Reversible Computing
New Devices Technologies such as Memristive, Spintronic, and Superconducting
Analog, Optical, and Quantum Computing
Memory-Based Computing
Heterogeneous Computing

Equally important as incorporating new devices and architectures, papers also addressed the need to co-design software and machine learning to take full advantage of the new hardware….