NewsCompany newsIndustry News



A Brief Talk on Integrated Circuits Becoming First-Class Discipline

On October 8, the Ministry of Industry and Information Technology (MIIT) published a letter in response to the CPPCC's Proposal on Accelerating the Independent Development of Industrial Semiconductor Chip Technology and Industrialization. In the letter, MIIT pointed out that the situation of the core technology of the integrated circuit industry in China is still not fundamentally changed, and it is urgent to strengthen the key technical issues to ensure supply chain security and industrial security. In addition to pointing out that the lagging development of industrial semiconductor materials, chips, devices and insulated gate bipolar transistor (IGBT) modules will restrict the transformation of new and old kinetic energy and industry in China under the current complex international situation, a series of industrial initiatives are put forward. In this reply letter, it is particularly pointed out that we should "promote the establishment of the First-Level Discipline of integrated circuits, and further do a strong demonstration. College of Electronics."

The establishment of the first-level discipline related to integrated circuits has been discussed for many years, but it can not be formed due to various controversies. With the continuous development and change of disciplines, the original division of first-level disciplines in fact has limited the training of integrated circuit talents in China, thus affecting the sound development of integrated circuit industry in China. Against the background of Sino-US trade frictions, this issue has been greatly magnified. The Ministry of Industry and Information Technology (MIIT), as the competent department of the industry, has a deep pain in the development of the industry. Therefore, the Ministry of Industry and Information Technology has been actively promoting the reform of integrated circuit personnel training. The statement also highlights the strong will of the ministries and commissions represented by the Ministry of Industry and Information Technology to promote the integrated circuit industry.

So why is it necessary to set up the First-Level Discipline of integrated circuits? What impact will the establishment of the first-level integrated circuit discipline have on personnel training and industrial development? How to build the First-Level Discipline of integrated circuits? The author of this article puts forward some opinions on his own experience and perspective for your reference.

First of all, we need to discuss why we need to establish the First-Level Discipline of integrated circuits. In fact, this is a huge problem, involving the education system, scientific research system, personnel training mode and other aspects of our country, but also related to the development and changes of the integrated circuit industry. The author will discuss it briefly first.

The first is the development of the IC industry itself. In Live: You may have learned fake digital integrated circuits, the author talked about the changes in the division of disciplines caused by industrial development.

Once integrated circuit is a kind of component, other components also include resistance, capacitance, inductance, magnetic devices and many other types. The circuit design uses different types of components to form functional circuits on the circuit board.

With the rapid development of integrated circuit technology driven by Moore's law, integrated circuit has developed from a component to the core of electronic system. Nowadays, most or even all functions of an electronic information system can be accomplished by an integrated circuit or several integrated circuits. Therefore, integrated circuit has changed from a component to the main carrier of electronic information system. This is also why the so-called "control" of the import of integrated circuits by the United States will have a great impact on China's electronic information industry as a whole.

It is precisely because of this technological progress that great changes have taken place in the design methods of electronic information systems. The design method of "building functional circuits with components on circuit boards" has been gradually abandoned and replaced by "chip is circuit" or even "chip is system". It can be said that the design of electronic system has changed into "making electronic system on chip". This has actually broken the traditional division of labor in the electronics industry.

However, the current division of disciplines of Electronic Science and technology in China is still based on the traditional model. Traditionally, the division of labour among the secondary disciplines of Electronic Science and technology is as follows:

MetInfo enterprise content manager system | MetInfo CMS

Microelectronics/Semiconductor is responsible for studying how to process materials into electronic components. The process of processing is "process". The subject of circuit and system is responsible for studying the "application of devices" and how to construct circuits with components. In this process, we need a process of circuit modeling and analysis. In fact, circuit analysis, analog circuit and so on are the core courses of our electronic information specialty. Communication, instrumentation, automatic control and other disciplines are based on the circuit to study how to form "complete equipment".

It may be a bit abstract to say so. Let's give a simple illustration with a less accurate example.

MetInfo enterprise content manager system | MetInfo CMS

This is an internal circuit diagram of an old SPC subscriber telephone switch found on the Internet. It can be seen that the internal structure of the old SPC subscriber telephone switch is composed of a motherboard with several extended circuit boards. If compared with the above division of labor, it can be roughly considered that microelectronics/semiconductor disciplines are to study those components on the circuit board, circuit and system disciplines are to study how to make these circuit boards, and the whole machine disciplines are to study how to use the circuit board to make this switch.

In 2010, I gave undergraduates a course called "Digital-Analog Hybrid Integrated Circuit Design", mainly about how to design ADC/DAC chips. At that time, an old gentleman came to attend the lecture. After listening, he and I discussed the content of the course carefully. At that time, I thought that Mr. Lao knew all the technical details very well. Later, I learned that Mr. Lao was the earliest batch of analog-to-digital converters in China at that time. But at that time, they designed a special signal acquisition board, which used many operational amplifiers and logic circuits to realize the functions of today's chip or even an IP core.

With the progress of integrated circuit and electronic information industry, great changes have taken place in the division of disciplines.

MetInfo enterprise content manager system | MetInfo CMS

The rapid development of integrated circuits has led to the monolithicization of electronic systems. The integrated circuit has been transformed from a component into the carrier of the whole electronic system. Therefore, the original research contents of circuit and system, even part of the whole machine discipline, are actually included in the scope of today's integrated circuit design. This has led to fundamental changes in the connotation and extension of integrated circuits. The study of integrated circuits is no longer a study of components, but a comprehensive subject from devices, circuits to systems and even software.

Let's see what the internal structure of the switch looks like now. The internal structure of a Huawei switch found on the Internet should not be the latest model, but it can also let us know the changes in it:

MetInfo enterprise content manager system | MetInfo CMS

It can be seen that in addition to the power supply circuit, there are also various components to match to constitute the circuit (which is in line with the current situation that the circuit and system discipline has essentially "retreated" to the specific circuit areas such as power electronics/microwave circuits), and there are only a few integrated circuits scattered in other large circuit spaces. The changes in the internal structure of these two switches reflect the profound influence of Technological Development on the design methods of electronic systems. However, the original division of disciplines has obviously been broken down and is no longer suitable for the development of new forms.

Then, according to the existing disciplines, microelectronics and solid-state electronics are only secondary disciplines under Electronic Science and technology. They are responsible for the research of electronic components (mainly semiconductor components, and other components are studied by other secondary disciplines respectively). The integrated circuit is only a three-level discipline under the second-level discipline, which obviously limits the development space of the integrated circuit profession. It's as if a child as old as eighteen is still wearing his clothes when he was five years old. It's bound to be frightened.

If there are many specific talks that will affect the talks, let me briefly talk about them here.

Firstly, the designation of talent training programs will result in the program being extremely inappropriate to the current technical system. Professor Yang Huazhong of Tsinghua University also mentioned this point in the interview. Professor Yang Huazhong of Tsinghua University: Successful IC entrepreneurs have these five characteristics

Professor Yang pointed out that the curriculum system of microelectronics specialty in China is not complete. Mainly reflected in our country's microelectronics specialty basically did not offer computer architecture, communication and network, media processing and other courses. For example, the main driving force for the development of integrated circuit industry has shifted from PC to mobile communication, network and media processing since 2000, but the domestic talent training has not kept up with it.

Strictly speaking, this pot should not be taught by microelectronics professionals. Because if we regard microelectronics as a major to train "component manufacturing" talents, the current curriculum system is still qualified. But the problem is that after the expansion of the connotation and extension of integrated circuits, we still regard integrated circuits as a branch of microelectronics specialty in the division of disciplines, and cultivate a large area of integrated microelectronics specialty curriculum training system for students. Naturally, many courses are not offered. However, in fact, this can not be solved by opening new courses in microelectronics specialty, but by redesigning the curriculum system in line with the existing training of integrated circuit professionals.

Since 2002, that is the year when I studied as an undergraduate. Some colleges and universities across the country have begun to offer the major of "Integrated Circuit Design and Integrated Systems" on a large scale. Later, with the construction of "Demonstrative Microelectronics College", the construction of this major has entered a new stage. However, due to the lack of independent disciplines, integrated circuit design and integration system specialty can not escape the constraints of the original curriculum and training program of microelectronics specialty. For example, in view of the problem of students'future graduate education, we have to reduce the class hours of "Microprocessor System Architecture" (former Microcomputer Principle), which is of great importance to IC design, and strengthen the class hours of XX Physics.

Above all, the influence on undergraduate training is greater than that on postgraduate training. Because the enrollment of graduate students is carried out according to the discipline, the evaluation and appointment of graduate tutors are also carried out according to the discipline. In the foregoing, the concept of components that have been broken through in integrated circuits has included many other subjects in the original subject division in terms of connotation and extension. However, due to the lack of independent integrated circuit disciplines, this leads to the fact that teachers engaged in integrated circuit direction research are scattered in many disciplines. Microelectronics/electronics science and technology account for a large part of them, but there are still many teachers in such disciplines as communications, computers, instrumentation and even automatic control. But because these teachers are not the mainstream of these disciplines, they have been squeezed or even suppressed to a certain extent in terms of job title evaluation and postgraduate quota allocation.

In addition, the subject construction in China is evaluated according to the way of subject evaluation, and the scientific research evaluation system in China has been seriously affected by the "Five Ways" for many years. Although the state has begun to pay attention to and rectify this problem since last year, it is still unable to effectively eliminate this impact until a new and effective evaluation system is established. For the research field of integrated circuits, the efficiency of issuing papers or collecting funds is obviously not high. In the absence of independent disciplines, integrated circuit teachers are inherently disadvantaged in all aspects of evaluation. The author used to say half jokingly and half seriously that if we use the criterion of "paper-only", we are "looking for death" in the research of integrated circuits. Take the direction of signal processing as an example, many times it is enough to publish a paper to write hundreds or even more lines of code and run several curves with matlab. If we continue to turn this algorithm into a chip (or at least design the prototype of the chip), it will take at least six months to one year. Finally, you will find that the "impact factor" and "partition" of the papers published on the chip are not as high as that of the previous pure algorithm. Under the traction of this baton, naturally many teachers have to change careers in order to survive. There are many stories in this area, but they will not be spread out one by one with limited space.

Of course, in order to solve these problems fundamentally, in the final analysis, we still need to establish a more perfect talent training system, evaluation system and other institutional construction and reform. But at this stage, it is impossible to build an ideal system in one step. By setting up integrated circuits as a first-level discipline, the personnel training program of integrated circuits can be formulated relatively independently, and teachers engaged in integrated circuit research can rely on it. This should be the most realistic way to solve the bottleneck of IC personnel training and technology research in China.

By building the First-Level Discipline of integrated circuits, we can straighten out the relationship between the current discipline of integrated circuits and other related disciplines, and make it conform to the development of today's technology. According to the characteristics of IC personnel training, we can reasonably plan the undergraduate training program and course content relatively independent. We can set up a more suitable evaluation system and postgraduate enrollment and training mode for integrated circuit disciplines. This will lay a good foundation for alleviating the predicament of our country in the training of integrated circuit talents. This is also of great significance in promoting the establishment of the First-Level Discipline of integrated circuits.

Of course, the establishment of the First-Level Discipline of integrated circuits has not yet been completely settled. After integrated circuits become a first-class discipline, it is still only a starting point for construction. Only when we can really build this subject well and play its due role can we meet the initial intention of setting up this subject. It is believed that many experts naturally have their own opinions on how to construct this subject. The author here dares to say a few words of his own ideas, the right to make a contribution to attract jade.

First of all, we need to straighten out all sub-disciplines/sub-disciplines within integrated circuit disciplines, and clarify their own positioning and tasks. Based on the analysis of the subject development in the above pictures, the author suggests that we can refer to the method of dividing the key projects of the National Key R&D Program "Photoelectronics and Microelectronics Devices and Integration". The subject of integrated circuit is divided into three sub-directions: integrated circuit and system chip, integrated circuit design methodology, device and process technology.

Secondly, it is necessary to set up reasonable undergraduate and postgraduate courses according to the three sub-directions within the delimited disciplines, so as to cultivate talents who meet the actual needs of integrated circuits in an all-round way. For example, in the direction of integrated circuit design methodology, the integration of mathematics, computer algorithms and microelectronic device technology should be strengthened to enable students to write simulation/simulation software.

Finally, and most importantly. It is necessary to introduce and train teachers in a balanced and reasonable way according to the scope of the subject. In order to ensure the sound and balanced development of disciplines, from a long-term point of view for the country's strategic needs stable delivery of talent.