Researchers at the Massachusetts Institute of Technology (MIT), working with chip maker Analog Devices, have successfully created the world's first fully-functional, programmable 16-bit processor to be constructed from carbon nanotubes - and they have used the open RISC-V instruction set architecture (ISA) to do so.
The march of Moore's Law - the observation turned ironclad commandment that the number of transistors on a leading-edge semiconductor part trends towards a doubling every 18 months - is not easy. As the number of transistors grow, the size of the components must shrink - and as we drop to 10nm and below physics begins to get in the way, as Intel and GlobalFoundries can attest. One area of research for improving matters is a switch away from silicon semiconductors to other materials, with carbon nanotubes one of the relatively exotic substances being tabled as a replacement for the silicon chip.
Back in 2013 Stanford University researchers produced the first carbon nanotube processor, using a 1,000nm process node and packing just 142 transistors into the test chip - enough to implement a small subset of the MIPS instruction set architecture (ISA) and create what they claimed to be the first Turing-complete general-purpose carbon nanotube computer.
Now, MIT and Analog Devices have announced the result of their own carbon nanotube semiconductor efforts: A rather more impressive 16-bit chip with 14,000 carbon nanotube field-effect transistors (CNFET) implementing the 16-bit version of the open RISC-V instruction set architecture.
'This is by far the most advanced chip made from any emerging nanotechnology that is promising for high-performance and energy-efficient computing,' claims Professor Max M. Shulaker of his team's creation. 'There are limits to silicon. If we want to continue to have gains in computing, carbon nanotubes represent one of the most promising ways to overcome those limits. [This paper] completely re-invents how we build chips with carbon nanotubes.'
The team's research processor was produced using a new technique, dubbed Designing Resiliency Against Metallic CNTs (DREAM), which greatly reduced the number of defects occurring during production, an issue which had previously prevented large-scale CNFET production. With its RISC-V ISA, the team's RV16XNano chip is capable of doing anything a commercial silicon microprocessor can do - albeit considerably more slowly, for now. As proof, the team had the chip introduce itself with a printed message: 'Hello, World! I am RV16XNano, made from CNTs.'
As with many post-silicon research project, the team is currently silent on a timescale for commercialisation - though has already signed an agreement with the US Defence Advanced Research Projects Agency (DARPA) to implement DREAM into a silicon chip foundry. 'We think it's no longer a question of if,' claims Professor Shulaker on the topic of CNFET commercialisation, 'but when.'
The team's work has been published as the paper Modern Microprocessor Built from Complementary Carbon Nanotube Transistors in the journal Nature.
November 22 2019 | 13:00