Today’s computers make use of one chip for computing and another chip for storing data. With the ever-growing amounts of data that need to be analyzed, however, the limitation of this set-up is becoming increasingly evident as the processing and storage capabilities of these computer chips have virtually reached their maximum. It goes without saying that more data simply means a slower processing time.
This is the problem that a team of researchers from Stanford and MIT are trying to solve and seem to have found a solution for. Instead of two separate silicon-based chips, they have developed a prototype that incorporates the processor and RAM into one graphene-based chip.
More specifically, the chip is made of two million carbon nanotube transistors — 2D graphene sheets formed into nanocylinders — for processing, with layers of resistive RAM (RRAM) placed over it, and ultra-dense wires inserted between the RRAM layers. Together, this makes up a ‘3-D computer architecture with interleaving layers of logic and memory’ that promises to address the communication ‘bottleneck’ being faced by present-day chips, and is probably the ‘most complex nanoelectronic system ever made with emerging nanotechnologies.’
As explained by Max Shulaker, lead author of the project, such kind of architecture cannot be used with today’s silicon-based technology because silicon transistors require extremely high temperatures of more than 1,000 degrees Celsius. A second layer requiring the same degree of temperature can’t be placed on top of that because it will just damage the bottom layer.
With their prototype, it’s completely different because both carbon nanotube transistors and RRAM can be made at temperatures that are much lower than those required by silicon, below 200 degrees Celsius only. This means a vertical arrangement or putting a layer on top of another won’t be damaging to the lower layer.
To demonstrate their work, the team had their carbon nanotubes act as sensors which they used to detect and classify ambient gases. They did this by placing more than a million carbon nanotube-based sensors on top of the chip.
Because of the layering of sensing, data storage, and computing, the chip was able to generate an impressive bandwidth, measuring each of the sensors in parallel and writing directly into its memory.
There’s a lot of potential riding on this new technology. For starters, both the carbon nanotubes and RRAM are more energy-efficient than silicon-based chips. Also, the integration of two chips into one eliminates the problems arising from moving data from one chip to another. Accordingly, the chip can store more data and process more of it — transforming large amounts into useful information within a reasonable amount of time.
Right now, the team is working on improving the fundamental nanotechnologies as they are exploring their new 3D computer architecture. The next step will most likely involve working with semiconductor company Analog Devices so new versions of the system — possibly one that integrates sensing and data processing into one chip — can be developed. As an example given, a device with such a chip could be used to detect signs of a disease simply by sensing specific compounds in a person’s breath.
In conclusion, Shulaker said: “The technology could not only improve traditional computing, but it also opens up a whole new range of applications that we can target. My students are now investigating how we can produce chips that do more than just computing.”
Details about the new chip were recently published in the journal Nature.