The further miniaturization of chips has encountered more and more technical limitations, and the ceiling of integrated circuit technology is looming. With the trend of future device miniaturization and high integration, the integration of computing and storage has begun. The value of memristors also increases.

The component revolution in the electronics industry

The memristor can be applied to a wide range of industries, but its most basic application is non-volatile impedance memory (RRAM). Compared with the current mainstream dynamic random access memory (DRAM) and flash memory (Flash Memory), memristive The energy consumption of the device is lower, and data can be saved even after the power is off. Specifically, whether the user shuts down or the global power failure, as long as the user turns on the computer next time, it can return to the previous state.

Some scientists are trying to use memristors to construct hardware with bionic brain functions to realize the hardware's "learning" and "associative memory" functions, and some scientists are discussing the possibility of the circuit itself adjusting its own state in real-time to meet computing requirements.

Combined with the memory capacity of the memristor, it means that the arithmetic circuit and the memory circuit can coexist at the same time and can be adjusted as needed. This has completely exceeded the design logic of this generation of computers, and many people hope that it will be applied to the research and development of smarter robots.

Experts believe that by 2020 at the latest, Moore's Law, which has witnessed the evolution of the semiconductor industry for half a century, will usher in the physical limit test. If graphene is a material revolution in the electronics industry, then the memristor is an electronics industry. The component revolution.

Significant growth in the huge demand market

Because of the advantages of memristors, such as small size, no power consumption to store data, fast read and write speed, long life, 3D design capabilities, and suitable for portable electronic equipment, industrial robots, supercomputers, servers and data centers.

The memristor market is expected to achieve rapid growth at a compound annual growth rate (CAGR) of 64.5% between 2018 and 2023.

Major companies in this field, such as Hewlett-Packard and Lenovo (the two have 28.66% and 17% of the supercomputer market in 2017) are developing this technology. With the constant focus on innovation in this field, memristors will be adopted by the market at a very steady rate.

Moreover, with the increasing demand for automation in multiple industries, the demand for memristors is expected to show an upward trend.

The rising demand for IoT applications and the deployment of artificial intelligence will also provide plenty of opportunities in the near future. However, the high cost associated with these devices, as well as the complexity in manufacturing, will challenge the growth of the market.

Memristors have emerged as a reliable and energy-efficient alternative to flash memory, so they can find applications in many industries. The increasing demand for automated processes in multiple industries will support the demand for industrial robots. These robots need to be able to provide zero start-up time, long life, and smaller flash memory.

Memristor is an emerging technology that can learn and store, etc., and provide brain-like calculations. The non-linearity of these devices makes them attractive for more complex brain-like controlled robots. These devices can work with extremely low power consumption, making memristor-based circuits more energy efficient than traditional transistor circuits. Therefore, due to the increasing demand for industrial robots, the memristor market will provide significant growth in the next 5 years.

Application and distribution of memristors

The memristor's storage and computing integration mode avoids the need to transfer the calculation results to the memory or external memory through the bus at each step in the traditional architecture for storage, thereby effectively reducing the load of frequent data access and transmission. Reduce the power consumption of information processing and improve the efficiency of information processing.

It has a potentially huge impact on modern electronics and can be applied to important fields of electronic information such as cloud storage, Internet of Things, consumer electronics, aerospace, earth resource information, scientific computing, medical imaging, and life sciences.

Because North America has the most advanced industries, ongoing research and development activities, and a highly information-based consumer base, North America will occupy the largest market share compared with other regions, and it will occupy the largest market share in memristor applications, including flexible electronics, Internet of Things and industrial robots.

In 2016, the installation of robots in the United States reached a growth rate of 14%, reaching a new record of 31,400. Moreover, major companies in this field, such as SanDisk and Hewlett-Packard, are located in the United States, which is an important factor that helps the United States gain competitiveness beyond other regions in the world.

Companies in the memristor field include Toshiba, Fujitsu, Sharp, Samsung, SK Hynix, Texas Instruments, IBM, Intel, AMD, Adesto, Cypress, HP, Winbond Electronics in Taiwan, China.

The subdivision of memristors

●Memory: The non-volatile memory capacity is the characteristic of the memristor, so the memristor plays the most significant role in the memory. The memristor is a basic storage unit with relatively small size and power consumption, which is not available in traditional units. Reached.

●Analog circuit: Traditional electronic devices such as diodes, transistors, thermistors, etc. do not have the characteristics of a memristor. These electronic devices combine with memristors to form a new type of hybrid circuit, and some of the analog circuits also appear. These new functions and new features make the circuit function more advanced and realize more circuit functions.

●Artificial intelligence computer: The memristor can also perform logical operations. This function can combine the data processing module with the storage circuit module, thereby changing the traditional computer architecture and bringing about the reorganization of the computer architecture. The development has brought new impetus.

●Analog neural network: In the analog neural network, the synapse function can generally be completed by software and hardware. Compared with hardware, the speed of software completion is slower and the overall efficiency is lower, so it is rarely used. The hardware is generally completed by analog circuits, analog-digital hybrid circuits and digital circuits.

●Computer chips: Memristors show more important potential in the field of neuromorphic computing chips. Neuromorphic computing is a new computing method that greatly improves computing power and energy efficiency by imitating the structure of the human brain.

Related post: Memristors introduction