Studies on the human brain, all of its functions and mysteries still unsolved, continue. Scientists are trying to adapt our brain, which has an enormous working mechanism, to technology.
Within the scope of these studies, scientists have been working on new generation computer chips modeled from the neural network of the brain for a while. If we look at the incoming information, it may be possible that the new generation computer chips modeled from our brain will be released before the end of this decade, thanks to a new material developed.
New technology could play a fundamental role in the development of next-generation computers
To put it simply, for this new technology, which is the first electrochemical 3-terminal transistor produced with two-dimensional materials, scientists from the KTH Royal Institute of Technology and Stanford University in Stockholm said that memory components produced with a titanium carbide compound called MXene “are an extraordinary way to complement classical transistor technology. states “potential”.
The findings, published in the journal “Advanced Functional Materials,” suggest that the new material, called MXene, could play a fundamental role in the development of neuromorphic computers similar to the human brain and thousands of times more energy efficient than today’s conventional computers.
Electrochemical random access memory, or ECRAM for short, acts like a synaptic cell in an artificial network, acting as a one-stop shop for data storage and processing. “These new computers will be based on components that can have multiple states and perform in-memory computation,” KTH Associate Professor and lead author Max Hamedi notes in a statement on the new technology.
This technology could be as fast as transistors in today’s computer
In a statement to TechRadar Pro, Max Hamedi explains that the technology uses the same processes as CMOS wafer assembly, integrating layers of 2D material onto silicon, so it’s a true hybrid integration with back-of-the-line processes.
Stating that the new technology has a faster write speed than other ECRAMs and that 2D ECRAMs are scaled to nano size, they can be as fast as transistors in today’s computers (sub-nanosecond); This means that the new technology can be integrated into existing computers using CMOS technology, thanks to the compatibility of the 2D metal of the transducer.
In this way, within 5 to 10 years, special purpose computer blocks that combine memory and transistors and make it at least 1000 times more energy efficient than the best computers used in simulation tasks with artificial intelligence today, can be produced, according to some calculations, even using this technology for certain algorithms, Hamedi said. states that it exhibits a million times energy efficiency. The launch of the first commercial computer chip modeled from a brain neural network can be expected by the end of the next decade, as a market-to-market strategy (GTM) requires at least 5 years of trials.
19471
