Geekly Articles each Day
📜 ⬆️ ⬇️

50 years of Moore's law


April 19 on Habré and Geektimes went unnoticed, although it was on this day that an article was published 50 years ago with some empirical observation, later called the “Moore's Law” , a sort of long-lived quick-moving and changing electronics world. Intel, in connection with this large and round date, interviewed its founder, old man Moore. Well, welcome to the cat for some interesting details and, in fact, an interview.

Interview


- How did you come to the conclusion that the number of transistors in a single chip will increase twice every two years? Was it a guess or the result of observations?

- In the early 60s of XX century. We developed semiconductor technology for its subsequent industrial application. With the tools that were then at our disposal, it was very difficult. I served as director of research and development at Fairchild Semiconductor, led the process of technology optimization. I was approached by the editors of Electronics Magazine with a request to prepare an article for their 35th annual publication, which had to be told about what would happen in the semiconductor industry in the next 10 years. Therefore, I decided to analyze what has been achieved. I took several microcircuits and paid attention to the fact that at first we installed one transistor on them, then eight each, and in the newest two times more, that is 16. In the laboratory we worked on creating microchips from the 30th elements and considered the possibility of installing 2 times more transistors. I came to the conclusion that every year we increased the number of elements by 2 times. Therefore, I made a bold assumption and said that we will continue to increase the number of elements by 2 times every year and move from 60 elements to 60 thousand within 10 years.

I wanted to convey to the public the idea that this is how the microelectronic industry will develop and this will have a positive economic effect. Although at that time there were few reasons for optimism: the first integrated circuits cost more than the individual body components for building similar circuits. It was necessary to tell that we have the technology that will seriously reduce the cost of microelectronics. I figured this was just a temporary industry trend. In fact, she was much more resistant than I could have imagined. One of my colleagues - it seems that Carver Mead, a professor at the California Institute of Technology - called the described trend Moore's law.
')
- Could you suggest 50 years ago that your observation will remain relevant for so long? This is not a law of nature, but a rule, which is the result of observations. So, there is no guarantee that it will continue to operate in the future.

- In 1965 and later, in 1975, when I clarified my observation, I did not think about when the law would cease to operate. The modern industry is finding new and new ways to make chips more and more complex. Today it is hard for me to believe that we are talking about billions, and not about tens, hundreds or thousands of transistors in the case of one chip. The law turned out to be more tenacious than I could have guessed in 1965 or 1975, and it is hard to say when its action will end.

Initially, the law was used simply as an illustration of the process. But gradually it began to be perceived as a defining trend of the industry, usually to measure progress. Now it is the basis of the development of the semiconductor industry.

- You have been involved in the development of modern computer technologies since their inception. What developments and devices have impressed you the most?

- In 1965, in the article I made many predictions, starting with watches and personal computers and ending with radar stations with a phased antenna array. Rereading this publication today, I myself am surprised at how accurate these assumptions turned out to be. Perhaps, I was most surprised by the evolution of the Internet and the place that it occupied in the lives of people all over the world. We knew that computers do useful things, that this will develop, but I did not realize that they would become the most important components of the global communication environment. I find it difficult to name another equally outstanding innovation.

- You once mentioned that you became an entrepreneur purely by chance. Now that you have a lot of experience with Shockley *, Fairchild * and Intel, what advice would you give to a tech entrepreneur?

- I'm not sure I can give useful advice in this area. I have always believed that it is necessary to determine the scope of activities and the type of products you want to produce, and then, if there is a practical sense in this, start. Now, many entrepreneurs approach this issue differently: they first decide to start a company, and then start searching for an idea that they could implement. Some of these projects have been very successful, for example, Google, but many could not stay on the market. I would advise you to pay attention to how to make the project long-term.

- Could Intel take its current positions and keep them if it hadn't followed Moore's law?

- I suppose not. When I am asked what I could say to Intel employees, I say that it is necessary to continue implementing the law. There are many problems that need to be solved in order for technologies to develop in accordance with a given speed, and this is very important because it allows you to control the situation and maintain success in the future. It is necessary to do everything possible, and this requires constant effort.

- What is the value of education for the technology industry?

- I think that the constant influx of well-trained engineers and scientists is a source of vital energy for any company, including for Intel. And this will happen only if universities teach students what we and other manufacturing companies need. Good ideas are often born in the minds of young engineers. They want and can make a real breakthrough that will create fundamentally new products and processes. Therefore, universities play an important role in supporting the success of modern technologies, and companies are becoming an important factor in the demand for and success of the learning process.



- Did Moore's law help Intel and other manufacturers to create innovations faster and to master the release of new products?

- Of course, the influence of Moore's law has changed over time. It was originally used simply to track progress. People created more complex chips. It was possible to compare the results and make sure that the complexity of the design is growing. But later, companies began to perceive the law as something that must be adhered to in order to remain competitive. In order to remain in the forefront, they need to move as quickly as indicated in the law. Thus, the law has ceased to be a means of assessing success and is now regarded as the main driving force of the industry.

- Moore's law is probably what made you go down in history. Or would you like people to remember you as one of the founders of Intel?

- It is difficult to name anything one, which I am proud of. Perhaps Intel creation is one of my greatest successes.

- How do you think, how famous is your law? How many people out of 100 will call you, if they are asked, what known laws in the field of electronics do they know?

- It surprises me how often in my life I come across references to this law. I checked the number of references to Moore's Law and Murphy's Law on the Internet. Google found at least 2 times more references to my law.



Examples to understand the "Moore's Law" on the fingers


Intel also prepared interesting comparisons and “applications” of Moore's law in everyday life.

For example, cars. If Moore's law existed in the automobile industry, then ...

In the construction could achieve incredible, such as:

Fly to the moon for a ridiculous cost, if Moore's law rules in aeronautics. For example:

Oddly enough, Moore’s law is not a law of nature. It retains its relevance, thanks to a whole army of people who develop the fundamental laws of physics. Now Intel factories produce more than 10 billion transistors every second, which are used in multifunction devices. And if earlier the first semiconductor transistors were the size of an eraser at the end of a pencil. Thanks to Moore's law, more than 6 million modern transistors (Tri-Gate) can be placed at the point at the end of a sentence.

Immediate perspectives and the future of Moore's law




At the moment, photolithography is carried out using a laser with a wavelength of 193 nm. This is some gold standard in the electronics industry. And with the help of various refinements, it is possible to achieve amazing results, for example, as the chip shown in the figure above, made according to the 14nm process technology. However, the era of EUV lithography (extreme ultraviolet) with a wavelength of 13.5 nm is not far off. EUV-tools allow you to do more operations in a single pass compared to modern scanners. Although the technology continues to evolve, this happens rather slowly due to a number of difficulties, including a radiation source and lenses for focusing the laser. Plus, new installations, vacuum systems and various materials for masks and photoresist are required. In addition, there are other difficulties, for example, reducing the "unevenness of the line width" instead of reducing the "unevenness of the edge of the line."

And in the distant future, in order to preserve Moore's law, it is necessary to make the transition to electronic lithography. With the help of technology Multiple Electron-Beam Direct Write (MEBDW) sometime in the future, billions of rays will create transistors on a chip. By the way, one of the developers of this system, Mapper, in the recent past received funding from RosNano.

PS: Article prepared on the materials provided by Intel.

Sometimes briefly, and sometimes not so much about the news of science and technology, you can read on my Telegram channel - welcome;)

Source: https://habr.com/ru/post/374943/

More articles:


All Articles