Professionals explain how the chip industry seeks new breakthroughs

Introduction

"Integrated circuits are not something that can be developed everywhere. I once went to a place where the local leader said, we have made up our minds to develop integrated circuits and build an integrated circuit factory here. I said, I am afraid that is not possible because you don't have the money. When I said I didn't have the money, the person was very unhappy and immediately asked how did you know I didn't have the money? I said to the person next to me, I will give you 5 billion yuan, you make this thing happen for me! I told him that I am afraid I have to add another zero at the end."

"Some comrades are very worried, saying that we have bought so many chips, what if they don't sell them to us one day? Isn't this being controlled by others, etc., right? It is natural to have such concerns, but if we put ourselves in their shoes, what are they worried about as manufacturers and suppliers? They are also very worried. Once a foreign friend asked me, he said, you bought so many chips from us, what will we do if you stop buying them one day? Everyone smiled knowingly."

Directly addressing the current "pain points" of my country's chip industry and exploring the path of breakthrough and high-quality development of the chip industry, heavyweight expert Wei Shaojun will give you an in-depth interpretation of "How to achieve a breakthrough in high-quality development from the "chip"?"

1. Who created the chip miracle?

Integrated circuits are a type of chip that we use every day. For example, there are more than 300 integrated circuits used in our homes. When we repair some electrical appliances at home, you can see a lot of black blocks. What are these black blocks? They are what we call integrated circuits and chips.

There are a large number of basic components of integrated circuits, called transistors, which may be billions or even tens of billions. The principle of transistors is very simple, but it took humans a very long time to invent such transistors.

We know that the world's first electronic computer was invented in 1945 at the University of Pennsylvania in the United States. We used the so-called electron tube, which is about two centimeters in diameter and five or six centimeters high. When powered on, it will light up like a light bulb.

Such a computer used 17,500 electron tubes, which was a lot, but the reliability of the electron tubes was very poor. One tube would burn out every six minutes. What should we do if one burns out? We have to replace it.

When changing, some ladies in the computer room have to run to turn off the power, change a tube, and restart the computer. Such a computer is very inefficient, so we urgently need to find a component that can replace the tube.

In 1947, three scientists at Bell Labs in the United States invented a new component that we later called the transistor. These three scientists were named Shockley, Bardeen, and Brattain. These three scientists won the Nobel Prize in Physics in 1956.

After the invention of the transistor, we saw that it was much smaller than the vacuum tube we were familiar with, smaller than a soybean, or even as small as a sesame seed. It had very high reliability and a very fast response speed.

In 1954, Bell Laboratories in the United States used 800 transistors to build the world's first transistor computer. This computer was used for the B-52 heavy bomber. It consumed only 100 watts of power, and most importantly, its computing speed was very fast, reaching 1 million times per second.

Transistors are very good, but people are still wondering if they can be made smaller. Why? Even if you use so many transistors, there are still solder joints. The solder joints will be cold, and the reliability will be poor. So can we find something with better reliability? So later we had integrated circuits, which are the chips we are going to talk about today.

On September 12, 1958, a young engineer named Jack Kilby, who was then working at Texas Instruments in the United States, invented the theoretical model of integrated circuits.

In 1959, a man named Bob Noyce, who was working at Fairchild Semiconductor and later became the founder of Intel, invented the integrated circuit manufacturing method that we all use today - mask exposure and etching technology.

So what we are talking about today is actually the technology we are using that was invented sixty years ago, it’s just that we are constantly reducing the scale and precision of it today. The integrated circuits invented by these two scientists have had a huge impact on mankind.

42 years after the invention of the integrated circuit, Jack Kilby won the Nobel Prize in Physics in 2000. Unfortunately, Bob Noyce had passed away by then, so he did not receive the Nobel Prize.

In 1962, the International Business Machines Corporation, or IBM, began using integrated circuits to manufacture computers. In 1964, it released a series of six computers worldwide, named IBM 360. It was extremely powerful and could perform a variety of tasks such as scientific computing and transaction processing.

A few years later, a young scientist at Intel named Ted Hoff designed the world's first microprocessor, called Intel 4004.

When this microprocessor was first born, its origin was not that impressive. It was used in calculators. A Japanese company went to Intel and asked Intel to help design a chip.

So Intel worked very hard on it, and finally gave the design to a Japanese company called Biscon, which made a calculator.

In 1981, ten years later, IBM organized a team to go to Florida to develop a major product that has influenced the world and mankind to this day, the personal computer, which we later called the PC. At that time, it used Intel's 8088 microprocessor, which was actually very slow, but it was very remarkable at the time.

Therefore, the progress of integrated circuits and chips has continuously expanded from the original government applications to civilian applications. For example, we have expanded from military applications to general civilian applications, and from general, conventional market commercial applications to ordinary people's homes.

There is a famous Moore's Law in the chip field. Its general content is:

When the price remains unchanged, the number of components that can be accommodated on an integrated circuit will double approximately every 18-24 months, and performance will increase by 40%.

For more than half a century, the evolution of chip manufacturing technology has continued to verify this law, and the speed of continuous advancement has continued to drive the rapid development of information technology.

What is the current level of chip technology development? Will future development reach its limit? Will Moore's Law continue to be effective? How many years can the miracle of the chip industry continue?

2. How amazing is chip technology?

How amazing is today's chip technology? It is constantly shrinking. How small is it shrinking? We have already achieved 7 nanometers, and it is estimated that we will reach 5 nanometers next year or the year after.

What do you mean by nano? I have no idea what nano means. I really have no idea. For example, you can imagine how small it is. Have we ever seen our own red blood cells? Of course we haven't, but we all know that a drop of our blood is red, because red blood cells are red, and the blood is reflected red.

How big is the diameter of a red blood cell? The diameter of a red blood cell is 8 microns, or 8,000 nanometers. According to our current technology, for example, a chip manufactured using 14 nanometer technology is about 40 nanometers in size, so we can put 200 transistors on the diameter of a red blood cell.

So you can imagine, such a sophisticated thing, precisely because it is so small, we can integrate a large number of things on a single chip.

Everyone will definitely ask a question: if we continue to move forward to 5 nanometers and then to 3 nanometers, can we continue to move forward? We believe that a certain technology may stop at a certain point, but it does not mean that new technologies will not appear. Two years ago, German scientists invented a new device called a molecular-level transistor.

In the future, our mobile phones may become smaller and smaller, to a level that we cannot imagine today. Of course, this does not mean that the size will be smaller, but the size of the mobile phone chip will be smaller, and the functions will become more and more powerful.

But any technology has its limits. It is impossible to have no limits. So what are the limits of chips? One is the physical limit. It is too small. In fact, there is also the limit of power consumption. For example, we all have electric irons at home. The power density of electric irons is 5 watts per square centimeter. 5 watts is very small, but it is very hot. We absolutely dare not touch it directly with our hands.

But what about integrated circuit chips? The general chip is tens of watts per square centimeter, so the chips we see often have a heat sink on the back and a fan on top.

When our power density reaches more than 100 watts per square centimeter, wind is no longer enough and we have to switch to water cooling. Water needs to flow through the supercomputer, and cold water goes in here and becomes warm water out there.

This kind of heat consumption and thermal effect is very strong. If it is not controlled, by 2005, the temperature of our chips will reach the temperature of a nuclear reactor, and by 2010 it will probably reach the temperature of the surface of the sun. So is it possible to use something so hot? It is impossible.

So people came up with a solution. We need to find a way to reduce the power consumption and turn the original single-core into a dual-core.

Later, when it extended to mobile phones, a particularly interesting phenomenon emerged. When people went to buy mobile phones, the salesperson told them, "Buy this phone. This phone has quad-cores. The 4-core version has more powerful functions and is better than that one."