The world's first! Tsinghua team successfully developed a "memristor chip"!

Recently, the team of Professor Wu Huaqiang and Associate Professor Gao Bin from the School of Integrated Circuits at Tsinghua University developed the world's first fully integrated memristor memory-computing integrated chip that supports efficient on-chip learning (machine learning can be done directly on the hardware end) based on the memory-computing integrated computing paradigm. This has made a major breakthrough in the field of memristor memory-computing integrated chips that support on-chip learning, and is expected to promote the development of artificial intelligence, autonomous driving wearable devices and other fields. The relevant results were published online in the latest issue of Science.

The chip contains all the circuit modules necessary to support complete on-chip learning, and has successfully completed multiple on-chip incremental learning function verifications such as image classification, speech recognition, and control tasks, demonstrating high adaptability, high energy efficiency, high versatility, and high accuracy, effectively enhancing the learning adaptability of smart devices in actual application scenarios. Under the same task, the energy consumption of the chip to achieve on-chip learning is only 3% of that of the application-specific integrated circuit (ASIC) system under advanced technology, showing excellent energy efficiency advantages, and has great application potential to meet the high computing power requirements of the artificial intelligence era, providing an innovative development path for breaking through the energy efficiency bottleneck under the traditional von Neumann computing architecture.

About Memristors

Memristor, full name memory resistor. From these two words, we can roughly deduce its function. The first person to propose the concept of memristor was Chinese scientist Shaotang Cai, who was teaching at the University of California, Berkeley. The time was 1971. When studying the relationship between charge, current, voltage and magnetic flux, Professor Cai, who taught at the University of California, Berkeley, inferred that in addition to resistors, capacitors and inductors, there should be another component that represents the relationship between charge and magnetic flux. The effect of this component is that its resistance will change with the amount of current passing through it, and even if the current stops, its resistance will remain at the previous value until it receives a reverse current and is pushed back.

Simply put, a memristor is a nonlinear resistor with memory function. Its resistance value can be changed by controlling the change of current. If high resistance value is defined as "1" and low resistance value is defined as "0", this resistor can realize the function of storing data. In fact, it is a nonlinear resistor with memory function.

Using a common water pipe as an analogy, the current is the amount of water flowing through it, and the resistance is the thickness of the water pipe. When water flows in one direction, the water pipe will become thicker and thicker with the flow of water. If the water flow is turned off, the thickness of the water pipe will remain unchanged; conversely, when water flows in the opposite direction, the water pipe will become thinner and thinner. Because such a component can "remember" the previous amount of current, it is called a memristor.

Because of their small size and low energy consumption, memristors are very good at storing and processing information. The workload of a memristor is equivalent to the utility of a dozen transistors in a CPU chip.

Comprehensive sources: Science and Technology Daily, Science Encyclopedia