Max Shulaker, assistant professor of electrical engineering and computer science at MIT, believes that we are only “three to five years” away from processors that utilise both silicon and the “wonder material” carbon nanotubes.
Shulaker is a member of the team that worked on building the first carbon nanotube processor revealed a few months ago: a 14,000 transistor strong CNT chip running on the open-source RISC-V instruction set. While still billions of transistors off today’s silicon CPUs, and a little sluggish in comparison, the researchers are already pushing to integrate their findings into foundries in the US. They’re also working with the industry to see how carbon nanotubes may help in overcoming the challenges facing silicon manufacturing, which will soon face rocky ground as compute demands skyrocket.
“We’re working with several industry partners to transfer this technology into the real world,” Shulaker says to PCGamer in issue 337. “We lead a very large programme sponsored by DARPA [Defense Advanced Research Projects Agency] who are already integrating carbon nanotubes into US Foundry, a company that does nothing but make chips.
“I really believe we’re a matter of three to five years away from having chips inside things you interact with that still have silicon inside of them, but also carbon nanotubes.”
Shulaker believes the key to carbon nanotube’s introduction into real-world applications is through layering them right on top of traditional silicon, atop of other layers in a 3D configuration, with a view to utilise their energy efficiency in low-power devices and to allow for greater performance on the next decade’s top chips.
The team also believes its research will enable CNT CPUs compatible with the x86 instruction set much beloved by Intel and AMD processors in the best gaming CPUs today.
Carbon nanotubes have been theorised, and demonstrated in the lab, for quite some time. This isn’t the first time we’ve heard of imminent applications for this nascent wonder material either. However, with promising evidence of working chips, and major backing in the US, there’s a chance we may finally see some early uses for these chips before long. And if anyone’s going to do it, it may well be some of the best minds in the biz over at MIT.
Header image of AMD chip courtesy of Fritzchens Fritz, Flickr / CC-0