IEEE Rebooting Computing Initiative Prepares for the End of Moore’s Law

Printer-friendly versionPrinter-friendly versionSend by emailSend by email
Monday, May 8th, 2017
unnamed

Tom Conte, professor. of Computer Science at Georgia Institute of Technology and co-chair for the IEEE Rebooting Computing Initiative

Much has been made recently in the engineering world about the impending end of “Moore’s Law.” For most of us, though, it’s unclear as to what exactly Moore’s Law is and what its predicted end has to do with any of us.

What we do know is that, for decades, our electronic products – our phones and computers, for example – have become smaller, faster, and cheaper all the while containing more powerful features year after year. For many of us, we instinctively try to resist the urge to run out and buy the newest smartphone, as we know the next iteration will be “better” for the price.

This decline in cost per feature is for the most part due to transistors, semiconductor devices that are the basic building blocks of microchips. In general, the more transistors there are on a circuit, the more powerful the device. Since 1971, transistors have become smaller and smaller, allowing more of them to fit onto an integrated circuit – the number roughly doubling every two years. This is Moore’s Law. The problem is that the transistor can shrink only so much. And this will have real consequences for consumers.

As Tom Conte, a professor of Computer Science at Georgia Institute of Technology and co-chair for the IEEE Rebooting Computing Initiative (IEEE RC) explains in an email interview, “Moore's Law states that transistors get half as expensive every X years. Thus you can get more transistors for your money over time. This has driven the remarkable growth of the electronics industry. But where X used to be two years, it's drifting towards three or more years. This means that the electronics industry – and all the consumer advances we've come to expect – is going to slow down.”

In terms of disruptive technologies, the slowing of Moore’s Law and the dramatic shift in how we compute will have profound effects, especially in the age of Industry 4.0.

So, how did we get there? According to Prof. Conte, we’re here “because of all of the decisions we made from the dawn of computing up to now. Not all of these decisions might have been wise.” The Rebooting Computing Initiative aims to usher in a new way of computing, but speeding up single thread performance would require the gargantuan task of completely rewriting the existing software. So, in 2012, he and Elie Track [co-chair, CEO of nVizix LLC and President of the IEEE Council on Superconductivity] decided to look into fundamentally different ways to compute, from new algorithms to new non-CMOS transistor circuits.

This ambitious decision was encapsulated in nine white papers published this year by the International Roadmap for Devices and Systems (IRDS). Supported by the IEEE Rebooting Computing (RC) Initiative, the IRDS white papers identify the challenges faced by the computing industry and make recommendations to guide the industry beyond the limitations of Moore’s Law. More importantly, they offer solutions to boost computing performance while simultaneously maintaining consumer affordability by leveraging new processing power, energy efficiency, and technologies. In short, the initiative’s primary goal is to seek out cost-effective replacements for the current transistors; the white papers were published to outline how the experts are tracking that, what the roadblocks to advancement are, and how the industry might get past these roadblocks.

The challenge is in the execution. “Industry is inherently risk-adverse,” Prof. Conte says, which works fine when the underlying technology can be built upon from generation to generation, but “when there is a big change, such as the one that is coming, industries rely on R&D.” The problem – as usual – is money. “Industrial R&D has been cut to the bone,” he continues in this email, and “most R&D is now done in national labs or academia. Moving the research from those entities into practice is going to take a lot of investment. This is where governments can make a difference.” He continues, “coupled with all of this is the problem that we have invested billions in software, both in person-hours for the programmers and in the software industry at large. A new way to compute is going to put great strains on the software industry and on the educational pipeline that trains our software engineers.”

Clearly, the end of Moore’s Law and the technical disruption to follow it will not only have far-reaching consequences for the field of engineering but will reverberate in the societies and economies of every country on earth.

This disruption, Prof. Conte points out, presents a real challenge to governments around the globe. Even the United States can no longer assume Silicon Valley will retain its central position in the computer industry.

Indeed, he advises, “each country must decide how to position itself to be a leader in the coming new age of computing” and recognize the opportunity to “steal the crown” from Silicon Valley.

 

 

Comments

samsung fire

new energy