Chapter 150 I can design, but you can’t build it (4k)
On the first day of work in 1960, Gao Zhendong received a notice from the 12th Machinery Department. One day later, he was invited to go to the 12th Machinery Department. The department's discussion is called communication on the next step in computer development.
After receiving this notice, Gao Zhendong decided to do nothing else today but one thing: take a good look back at the development of computers in the next few decades and put their appropriate content in the form of planning suggestions. , introduced at the meeting.
This matter does not need to be drawn by lottery or wasting search opportunities. If it does not involve technical details, the development history of computer technology that Gao Zhendong has been exposed to can make a big difference.
After all, in principle, since the invention of semiconductor technology, there is no difference in core principles between computers sixty years later and computers today.
As long as we grasp the development of the core few things, we will not be left too far behind. With Gao Zhendong here, we may still be ahead in individual projects.
After thinking about it, Gao Zhendong wrote a few words on the paper.
——"Computer Technology Development Plan and Suggestions"
Gao Zhendong was thinking and writing. He was not going to come up with a large and comprehensive planning suggestion at once. Many technologies are It will not be produced and applied until the technical conditions develop to that point.
Gao Zhendong’s mentor in his previous life once participated in the preparation of similar plans. In the words of his mentor, planning must be forward-looking, advanced, comprehensive, and complete, but one thing that is taboo is greed. Being greedy for the big and seeking perfection, wanting everything, ends up spreading limited resources, resulting in nothing being done well.
After all, there is no "Showmethemoney" or "Operationcwal" in the real world. The realization of any plan requires the investment of resources and the accumulation of time.
Why do most countries, in addition to planning such as grand strategy, span a very long time period, there are very few countries that can actually do a good job in developing development plans that span the span of the "Five-Year Plan"? Not much.
The height of the Twelve Machinery Department is obviously not high enough to formulate strategies. For example, "taking information technology as one of the scientific development strategies" is not the turn of the Twelve Machinery Department. That's a higher level thing.
Therefore, the communication with the 12th Machinery Department was more at the level of tactical execution, so Gao Zhendong wanted to focus on some technical nodes or technical directions that were necessary, far-reaching, and significant.
Gao Zhendong was thinking about writing, and the time soon arrived in the afternoon.
Looking at the paper, there aren't actually many items left. Many of them were crossed out after being written down and found to be inappropriate.
For example, he wrote about the popular artificial intelligence in his previous life and then crossed it out.
The core mathematical principles of artificial intelligence, deep learning and other technologies have appeared very early, and there are some applications. However, it was not until 2010 that it became popular on a large scale, and it was not until 2020 that it became popular on the entire Internet. It is not without reason .
Gao Zhendong’s mentor once said that some companies introduced neural networks into water conservancy and hydrological forecasting calculations in 2000. After a large amount of data learning, it was still passable, but there were A small problem: slow!
For ordinary hydrological agencies without supercomputers, by the time it calculates the forecast results, the flood will have already passed a day or two.
Its actual application effect is far less good than using traditional methods or empirical formulas.
Correspondingly, if you want to engage in artificial intelligence at this time, you are undoubtedly thinking too much.
In the end, Gao Zhendong only left a few pieces of content closely related to equipment manufacturing and popularization.
As for the longer term, when it needs to be raised, Gao Zhendong's influence should be enough. Right now, it's not even close.
You can’t eat hot tofu in a hurry, and you can’t become fat after one bite.
Gao Zhendong, who felt that he was almost done, stopped writing and let himself take a rest.
This break reminded Gao Zhendong that today is Sunday. Although he is at work, he should not delay the accumulation of lottery draws.
Looking at the accumulated number of draws, sure enough, it was six.
Gao Zhendong decided to smoke once, and just hold it on his hand 5 times for emergency purposes.
His consciousness swept over the pile of materials in the storage library. When he saw the names of the materials, Gao Zhendong finally restrained his urge to print instead of handwriting.
Be patient, be patient, there is no need for it now, wait until things move forward.
——"Single Crystal Furnace Structural Design"
It should be something that can be found everywhere in major knowledge websites on the Internet, not only in the future.
Single crystal furnace, a device that uses the Czochralski method to grow single crystals.
Not only silicon single crystal, but also many kinds of semiconductor-related base single crystal materials can be produced with this thing.
For example, the mercury cadmium telluride material crystal used in the line scan infrared thermal imaging camera extracted earlier.
Moreover, the consistency of the materials produced in this way is very good. This is no nonsense. Can the consistency of single crystal be bad?
If I remember correctly, according to the history of domestic semiconductor development, the country was engaged in the drawing of silicon single crystals at this time, and it should have been successful in 60 years.
But even if it succeeds, if you want to compare with this mass production performance and crystal performance, it will definitely not be comparable. This thing in hand integrates mature technologies and concepts in the future, and the original predecessors now It's definitely not the same thing.
If nothing else, the crystal diameter is not on the same level.
Seniors working on silicon monocrystals, I'm sorry, I can't wait.
Early the next morning, Gao Zhendong rushed to the 12th Machine Department.
In the conference room, there were the leaders of the 12th Machinery Department, the leaders of the National Defense Work Committee, Director Li of the Computing Institute, Chief Engineer Chi, and several unknown people, who should be from semiconductor manufacturing plants.
The meeting did not have too many cumbersome procedures and went directly to the topic.
When he learned that the National Defense Work Committee and the 12th Ministry of Machinery needed to confirm the development time of the next generation of transistor computers, Gao Zhendong thought for a while.
"If we just talk about design, I can come up with the design of the next generation computer within one to two years. But when it comes to manufacturing, I think it will be more difficult with our current capabilities." Gao Zhendong is confident in saying this. , especially after the 4-for-1 search function. In his collection of materials, there are many IP core designs of classic CPUs, and there are also many detailed explanations of classic CPU architecture and design.
This is what people who work on CPLD/FPGA like to do, and some of them also like open source.
IP core design can be simply understood as the design drawing of the CPU.
The time Gao Zhendong said surprised everyone present, but the leader of the 12th Machinery Department was a little surprised that he couldn't make it up. Aren't they all transistors?
"Comrade Zhendong, aren't they all transistors? If you can build DJS-59, why is it so difficult to build your next-generation computer?"
Gao Zhendong talked eloquently: "Dear leaders, in my plan In my opinion, the operating frequency of the next generation computer is about ten times that of the current DJS-59, the word length is expanded to 16 bits, and the size remains the same or further reduced. I am very dissatisfied with the speed and word length of the current DJS-59. "I'm satisfied, but it's not good enough to use."
When they heard about the improvement span, which was like shooting a satellite into the sky, the people from the National Defense Work Committee, the 12th Machinery Department, and the Operations Office were both happy and surprised.
The good news is that Gao Zhendong never fires off cannons. Since he said so, he must be certain.
What is surprising is that Gao Zhendong's performance improvement plan is so radical. The span is a bit scary.
The leader of the 12th Machinery Department made a final decision: "Comrade Zhendong, if you really think so, don't dare to say anything else. Our 12th Machinery Department fully supports it."
The leader of the Defense Work Committee also did not hesitate: "Prevention The working committee system also fully supports you.”
Can you support it? Since the mass production of DJS-59, the benefits it has brought to related work are visible to the naked eye, and it is said to be ten times faster. Times, the word length has doubled, so we have to keep up with it.
Gao Zhendong shook his head: "It's not that simple. As the performance of the computer I plan to improve, the number of transistors will increase dramatically."
What are these words? Everyone nods in approval, and performance improvements must come at a price.
"The number of transistors in this computer, not counting the storage part, only counting the host body, I estimate that the number of transistors will be as high as 30,000 to 40,000. Coupled with the performance improvement and doubled word length, the transistor memory must also be increased accordingly. The overall increase in the number of transistors is a very large number.”
Hearing this number, everyone was a little silent. The price of this number of transistors is incredible.
The leader of the National Defense Work Committee smiled and said: "Comrade Zhendong, if the performance of the machine really reaches the level you mentioned, it would be acceptable to increase the price ten to twenty times."
He knew very well that this was simply a leap from quantitative change to qualitative change.
Others also nodded in agreement, but what they didn't expect was that the biggest problem was not the price.
Gao Zhendong shook his head: "It's not a price issue, but a reliability issue. At this scale, the reliability, availability, electromagnetic compatibility and other performance of a machine composed of tens of thousands or even hundreds of thousands of transistors will drop sharply, or even It can’t be used.”
It was only then that everyone realized that there was still this trouble. It seems that high-tech matters are not that easy to rely solely on quantity.
"Then you should have a solution? Since you have already planned the next generation of computers, I believe you must have your own considerations for the problems." Based on Gao Zhendong's performance at work Attitude, the leader of the 12th Machinery Department believed that he must have considered it and would not just ask questions without proposing a plan.
This is actually an issue that you should pay attention to when you report to your leadership and seek support. It’s okay to complain about difficulties, but don’t just offer support without offering any solutions. When talking about problems, you should bring some support. If there is a solution, the possibility of getting support is much greater.
As expected, Gao Zhendong nodded: "There are several ways, I will tell you one by one."
"First of all, I plan to use integrated circuits to solve the problem. Integrated circuits are a new technology abroad, and now they are gradually becoming practical. They can put many digital logic devices into one small component. . ”
“This can greatly simplify the circuit complexity of the computer and solve the problems caused by a large number of transistors.”
Gao Zhendong’s idea is to use integrated digital circuits to build the CPU. Not as good as a fully integrated CPU, but usable at this scale.
Just like digital logic circuit chips such as 74, 54, and CD4000, although these circuits are still in their infancy, in terms of time, it is almost enough.
Everyone present has heard about integrated circuits, especially the technicians who are more familiar with it.
People from the semiconductor production plant, in fact, people from the 1274 factory, asked: "But we don't have integrated circuit technology now. We have also talked about this matter with relevant research institutes, but we haven't found anything yet."
< br>The first domestic integrated circuit was originally manufactured in 1965.
Gao Zhendong nodded: "I know, so my idea is to walk on two legs. There should be some integrated circuits I need abroad now, and we can find a way to buy some."
< br>"At the same time, we can also carry out research on our own in China. If successful, we will start domestic replacement work."
"From now on, when my design is completed, there will still be some time in between. During this time Whether we wait for foreign technology to mature and more integrated circuits that better meet my requirements appear, or we develop domestically, we should be able to achieve some results and just keep up.”
Seeing that everyone was thinking, Gao Zhendong took advantage of the situation and put forward his own ideas: "Actually, I came here today just to make my own suggestions in this regard."
The leader of the 12th Machinery Department signaled to Gao Zhendong Zhendong continued.
“First of all, it is recommended to tilt resources towards integrated circuit manufacturing. From the perspective of the current technology development direction, information technology, including computer technology, is one of the main directions for future development. Information technology can go deep into production All aspects of life, like the steam engine and internal combustion engine back then, are one of the main technologies for the next productivity revolution. "
It is not said to be the "main driving force". At this time, it is better to use more refined words. .
“Semiconductor manufacturing technology is one of the core basic technologies of information technology. Without it, no matter how good we design something, we will not be able to produce it, and we will only be stuck by others.”
"Now that semiconductor technology has just started, it is when the gap between us and foreign countries is smallest. As long as we invest certain resources, we can at least keep up with foreign countries and not fall behind others. We can even make progress in some specific aspects." A little location advantage project allows us to take a little initiative in certain situations.”
These few words made several leaders present flash their eyes and seize the opportunities of emerging technologies to keep up with the international market. level, develop regional advantages, and maintain strategic initiative.
The two leaders of the National Defense Work Committee and the 12th Machinery Department exchanged glances. The leader of the 12th Machinery Department smiled and said: "'Information Technology', this is a good term. Ah. He is indeed a cutting-edge technical talent with many achievements. I didn’t expect that in addition to specific technical achievements, he also has great ideas in scientific research strategies. Do you have any specific suggestions?”
Gao Zhendong came up with what he wrote. The piece of paper was handed over: "This is my specific idea in this regard. It may not be mature. It is a personal opinion. It is for reference only."
(End of this chapter)