If you compare Moore's Law to one person, this year he is already a 50-year-old middle-aged. In the past 50 years, the electronics industry has made great progress following Moore's Law. However, today people will also wonder if Moore's Law is outdated, and it can still adapt to the times and continue to guide the development of the future electronics industry. In the last part of this article, we will return to the 1960s to see what kind of era background and practice Moore proposed Moore's Law, and found the other side of Moore's Law that was not taken seriously.
Half a century ago, a young engineer named Gordon Moore looked closely at the emerging industries he was in and made his own predictions about the major events that will take place in the industry over the next decade. The four-page forecasting article was published in Electronics magazine, a young engineer predicting home computers, mobile phones, and cars with automated driving systems in the future. In his article, he predicts that the number of electronic components that can be accommodated on an integrated chip will be doubling steadily each year, which will make the integrated chip more and more economical, and this is the technology that drives his whimsy into reality. The power.
Ten years after the article was published, the "Moore's Law" describing the exponential growth of electronic components on integrated chips has not stopped. Today, Moore's Law has been running through the rapid development of science and technology in the past 50 years. Modern technology that follows Moore's Law has brought computers, personal electronic devices and sensors to people's lives. The influence of Moore's Law on modern life is hard to estimate. Without the continuous development of integrated chips, we can't sit on the plane, can't make a phone call, and can't even use the dishwasher, let alone discover the Higgs boson and create the Internet. .
But in the end, how does Moore's Law affect our lives, and why can it achieve such an achievement? Moore stated that technological progress is unquestionable and overwhelming, or does it only reflect the state of technological development in a particular period? To this day, can we use Moore's Law to explain the progress of technology and innovation in computer science in the past decade?
In my opinion, the status of Moore's Law is unquestionable. Moore's Law is proof of the times, which represents the hard work, intelligence, and motivation of the free market. Moore's prophecy was initially a simple observation of an emerging industry, but over time it has become a self-fulfilling expectation. The realization of Moore's Law is the result of continuous creation by thousands of innovative companies and engineers who can see the potential of industry development from Moore's Law and do their best to maintain technological leadership, otherwise they will risk falling behind Risk to competitors.
But I also want to say that although Moore's Law is endlessly applied to explain various technological advances, it is not just a simple concept. The meaning of Moore's Law has undergone repeated changes in these decades of development, and it is still changing today. If we want to understand the nature of technological progress from Moore's Law and predict the future development from it, we need to have a deep understanding and observation.
In the 1960s, when Silicon Valley was still not famous, the young Gordon Moore was the R&D director at Fairchild Semiconductor. After leaving the Shockley Semiconductor Laboratory, he founded the company in 1957 with others. In this company, they completed the early development of silicon transistors.
Fairchild was one of the few companies developing silicon transistors at the time. Transistors are variable current switches that control the output current based on the input voltage and can be used to calculate and store data. Fairchild quickly discovered the niche market.
At the time, most circuits consisted of a single transistor, resistor, capacitor, and diode connections that were manually assembled on a single board. In 1959, Fairchild's Herni invented a planar transistor that replaced the previous mesa.
With this planar transistor, engineers can interconnect multiple transistor traces and mount them on a small or small piece of semiconductor wafer or dielectric substrate to create something called an "integrated circuit." Jack Kirby of Texas Instruments is a pioneer in integrated circuits. He first thought that resistors and capacitors (passive components) can be made of the same materials as transistors (active devices). Moore's colleague Robert? Neuss has shown in practice that planar transistors can be used to fabricate integrated circuits by attaching an insulating oxide coating to the transistor and then adding aluminum wires to connect different transistors. Fairchild has invested this new manufacturing process in the production of the first silicon integrated circuit, which was introduced in 1961 and has only begun to contain only four transistors. By 1965, the company had been able to produce integrated circuits containing 64 electronic components.
With the accumulation of these prior knowledge, Moore made a bold conclusion in a paper published in 1965: Integrated circuits represent the future direction of the electronics industry. This statement is of course self-evident today, but it was controversial at that time. Many people question Moore's point of view that integrated circuits are just a small branch of the electronics industry.
These doubts can be forgiven because the process of integrated circuits at the time was much more complicated and expensive than other manual board products - from today's computing point of view, the cost of integrated circuits was as high as $30 at the time. The cost of a single component is less than $10. At that time, there were only a handful of companies producing integrated circuits, and their real customers were only NASA and the US military.
But to complicate matters, the transistors at the time were not reliable. According to Moore, at the time, a single transistor could only perform about 10%-20%. When you integrate multiple transistors on the same board, although it is expected to perform its best, the effect is not satisfactory. This situation occurs because the operational logic is flawed. Although eight transistors are integrated on the same board, they do not actually perform as a whole, and the effect is equivalent to eight independent transistors. This is because the probability of each transistor failing is independent, and such failures occur randomly, such as splashed paint can cause the transistor to fail. If one of the two adjacent transistors fails, the two transistors will strike at the same time. Therefore, when two transistors are connected together, there is a risk of damage.
Despite all the difficulties, Moore still believes that integrated circuits will one day prove to be an economical choice. In his paper published in 1965, in order to prove that the integrated circuit will have a bright future, Moore constructed a logarithmic model by using Fairchild's first-generation planar transistors and a series of integrated circuits that were subsequently produced. In this model, he found that the number of components on an integrated circuit doubled each year over time.
By adding a small trend line to the model, Moore made a bold inference: this growth trend will last for 10 years. In 1975, he predicted that people would see the number of components on integrated circuits increase from 64 to 65,000. This prediction is quite close to reality. In 1975, Intel's charge-coupled (CCD) memory chip was ready to produce 32,000 components. After a year of development, the results would be quite close to Moore's prediction. This Intel company was established in 1968 after Moore and Neuss and Grove were separated from Fairchild.
Guangzhou Ehang Electronic Co., Ltd. , https://www.ehangmobile.com