According to the British “New Scientist” weekly website reported on July 1, the energy consumption of computers that use molecules to solve problems is one-ten-thousandth of that of traditional computers. If made bigger, these biocomputers can effectively solve complex logistics problems that typically require a lot of time and effort.
For most of the history of computers, as the size of the chip became smaller, so did the energy required to operate. But that relationship broke down about 15 years ago, which means that computers performing large computing tasks aren’t as energy efficient as we once hoped.
One way to make future computers more energy efficient might be to abandon electronics altogether and switch to biology. Thiel Koten and colleagues at the Technical University of Dresden in Germany built a chip-based biocomputer that uses molecules moving through channels to solve problems.
This chip is made of glass. To perform the computational task, the researchers filled the chip with a fluid containing a driver protein molecule and microtubules.
Microtubules form part of the “scaffold” inside the cell, and the driver proteins move along them to transport other molecules. But the design of the biocomputer reverses that situation. In a biocomputer, Koten said, microtubules moving in the channels of the chip actually “surf” on the driver protein. All microtubules move simultaneously, meaning that many computational tasks can be performed simultaneously.
The microtubules move through the channel, each path corresponding to a tentative solution for the computer. The researchers then took images to read the output information from the biocomputer and determine the most successful solution.
Coten said biocomputers can solve dense combination problems, such as calculating the best route for an airplane that must stay in multiple cities.
Henry Hess of Columbia University in the United States said that the new biocomputer mentioned above is a major advance compared to the first generation of biocomputers made a decade ago.
Traditional computers are able to solve specific problems faster, but to solve a combination problem that contains more variables, it can take billions of years for an electronic computer. Dan Nigulau of McGill University in Canada says how to increase the computational speed of electronic computers is an open question, and for new biocomputers, researchers only need to add more molecules to solve the problem in a matter of days. Microtubules are very small, and 1 gram of fluid can hold trillions of microtubules, which is an advantageous condition.
The “surf” molecules also use only one-tenth of the energy consumed by each step of the computation in traditional computers. “It’s understandable that these [molecular] motors have been optimized for 1 billion years in an evolutionary process,” Cotten said. ”
This new device is by far the most powerful of its kind, but it is not advanced enough in terms of practicality. To be practically applied to areas such as logistics, The team’s computer needs more molecules, which is a challenge for manufacturing technology, and microtubules “bending” as they move through the channel can cause more errors, Cotten said. “The biocomputers we build are in the state that electronic computers were in when people first started assembling the first transistors,” he said. ”