The amount of storage our everyday programs are occupying is increasing day by day. Take games, for instance. An average AAA PC game is well over 40GB these days and it continues to spiral upwards. We can't possibly imagine the size of AI programs and other complex software. Silicon chips can't effectively process this huge load of data so researchers at MIT and Stanford have come up with prototype carbon chip to tackle this problem.
Our everyday transistors currently use silicon and, to some extent, germanium is involved. Group IV elements are basically semiconductor materials but silicon is becoming more and more inefficient as we move into advanced computing. The alternative found out by researchers is Carbon Nanotube transistors which use 2D graphene and non-volatile memory (RRAM - Resistant Random Access Memory).
The goal was to combine logic and memory to create a 3D architecture which will address the communication bottleneck by stacking the RRAM and carbon nanotubes on top of each other in the form of layers. In other words, they are built over each other, providing a denser array. This is something which is not possible with traditional silicon chips since they are only utilised in 2D architectures.
Now the question arises, what is the prime benefit of these new chips? Well not only are the new chips more power efficient but they can also effectively compute a ginormous amount of data due to overcoming the bottleneck of moving data between chips. Traditionally this proves to be a time and energy waster.
The new 3D architecture also promises to continue Moore's Law which, until recently, was under question due to issues in the developing the ever-decreasing sizes of silicon transistors necessary. Scientists can't split the atom (at least not to help play Destiny 2 they can't), and there are theoretical limits to how small they can go.
However, this new architecture will allow a greater number of chips to be integrated per unit volume. A practical demonstration of carbon nanotube chips was given by the researchers in which they were made to act as sensors to detect ambient gases.
We are talking about a phenomenal change in the architecture of computing here. If you are interested in finding out more, then head over to MIT's research blog here.
What are your thoughts on this new 3-D architecture? How much time will it take for Silicon chips to be replaced? Let us know folks!