The new processor is a 32-bit core with a 16-bit external bus, which is exactly the same when it is produced, and if necessary, the external package is changed, and it is a new 32-bit processor.

Computer architecture is difficult to develop, just like martial arts, and it is a few people who can find a new way to start a sect, such as Turing, von Neumann, David Patterson (RISC), and John Hennessy (MIPS).

The long-term architecture is the big road goods in his memory, the pure-blooded RISC, the streamlined can not be more streamlined, and the design has been tempered by thousands of degrees, which is why there are so few transistors but the performance is outstanding.

Compared to the CISC processor, the core and design of the RISC processor are easier to find, because they are all open source, and even perverted like ARM, as long as you use his core design, everything will be given to you.

I have studied several classic architecture designs in my dream for a long time, and the 8-bit is practiced, because this is the clearest memory, and the 32-bit one has also been done, but it is much larger, and the recognized simplest structure has to be stuffed with more than 30,000 transistors.

The best processor in the market in later generations is the risc type is ARM, which is overwhelming, but this company does not produce, but only provides core design for licensing, and the system is streamlined and cheap, so it is the favorite of system integrators, and most of the monolithic systems use ARM cores, such as mobile phones.

It's just that this company focuses on system inheritance and has a poor reputation, while in the desktop field, IBM's power series has been flourishing for a long time, and Sun's SPARC is also very famous.

From the design to the manufacture of the processor, the investment is huge, and there is no extra money to pay for the tuition for a long time, so I have to carefully consider where my products will develop in the future, desktop or monolithic system.

Monolithic systems require a single volume, similar to electronic components, and the retail price is less than five US dollars. I don't want to produce this kind of goods in my own factory for a long time, these are customized projects, there are no millions of sales to talk about profits, but it is possible to authorize manufacturing in the future, so pass.

If you focus on the desktop system, the competition can be regarded as cruel, and most of the world's powerful companies are eyeing this area. The fierce market battle after the 90s is still clear in my memory for a long time, and it took a lot of thought to take this step.

Fortunately, for a long time, he is also a capable person, and he has not yet reaped the benefits of Intel, if he has so much experience in later generations, he will not regret his death?

What's more, with the help of strong people like Dr. Yu, it will be the pillar of Intel in the future.

With the desktop system as the goal and occupying the mainstream market, it is necessary to at least conquer Europe and formulate standard chips, which is the weight to compete with a group of subsidiaries in the United States, otherwise the consequences are unimaginable.

After all, what is produced is standardized, and it can be sold to all manufacturers, and the return is many times the same as the investment in a customized system, and the return is many times, and it is not dependent on one company, which is the king.

Time was running out, and a group of people began to retreat for a long time. I also consulted Dr. Yu for a long time, how Intel is engaged in the processor project, he is just a wild man, and he is not even as good as someone from the Chinese Academy of Sciences. In the past, 021 was a small trick, and it was okay to do it alone, but now it is unthinkable to develop a 32-bit CPU, and it is unimaginable without rules and plans, and the efficiency is immediately different.

Dr. Yu is also a person from the past, Intel's project management and quality control are very unique, which is unique, and has not fallen in decades of competition of the jungle, and these systems have a great effect.

Dr. Yu told Chang Chang about his time in Fairchild and compared it with Intel's experience. Fairy Children's talents are not many, and the leaders are all god-level figures, but the big boss Xiao Keli is mean and unkind, and his subordinates are very strict, which led to a collective mutiny among several of his generals and made Intel.

Later, Fairchild was acquired, and those who could make decisions were laymen. At that time, there was a saying in Silicon Valley that marketing management talents could make achievements in that industry.

There is some truth in this statement, but it is not reflected in Fairchild, and the electronic engineers jump away one after another, in vain.

In fact, it can't be regarded as the fault of the fairchild, this is the atmosphere at that time, as long as the results are made, the first thing who thinks is to find a venture capitalist and start his own business.

Intel also suffered losses in this regard, and the engineering fagan who developed the first microprocessor left to hurt Intel.

Maybe this is the charm of high-tech enterprises, as long as you have technology, money is not a problem, this is an era of heroes, and individual talent can support a dynasty.

But the old company with a microprocessor foundation is not the same, Intel is a whole, and the last thing that is afraid is that the project leader will leave. Anyone or group who seeks another job will always find a second echelon.

And Intel will never be stingy to give newcomers the opportunity to stand out, as long as you have the strength, you are the top, as long as you have potential, you have to bear the burden. Moreover, the newcomers have no baggage to do things, and the veterans are willing to help, so the chances of success are much greater than failure.

Therefore, Dr. Yu suggested that Changjiu recruit and build a team first, and Changjiu himself should be the project manager, rather than the chief engineer.

"I don't know how to manage, that project manager is a shrimp thing?"

"It's okay if you don't understand, you just need to remember a little bit. "The responsibility of the project manager is actually very simple, which is to ensure that the plans of other research members can be carried out smoothly, and you only need to grasp the general direction, and the specific difficulties are solved by others." Your design is so good that you can't draw it yourself, it's a waste. ”

Chang nodded: "It makes sense, I feel like I'm too busy to be busy, and I don't have the skills." ”

Dr. Yu smiled and said, "Be busy to the point, otherwise you will be blindly busy, and you will get half the result." You bring some project leaders, just discuss the progress of the research with them every day, and provide some suggestions. ”

His eyes widened for a long time: "Just say that you can talk, but what about the rest of the time?"

"Well, the rest of the time you just read books and do the research you want to do. You are good in English, here are a few books on management, take them and read more. Dr. Yu made a list of books, "I used to be a little bit unable to change, you must know that I was doing semiconductor physics research, and I was very confused as a manager for the first time, and then it was like this, talking and making suggestions, giving your subordinates a little encouragement, and my subordinates were much faster than you, and I did my own research in the extra time." Management doesn't require you to put too much effort into it. ”

Confused for a long time: "It sounds very laid-back, it feels like a mess." ”

"Of course, it's leisurely, because the production line hasn't opened yet, which is my experience at Fairchild. "But Intel's management is not so leisurely, for example, a team of three people, everyone has a heavy task, division of labor and cooperation, separate solutions to different problems, go hand in hand." The project manager is like a coachman, trying to control the direction of progress, and make all the horses work reasonably. ”

Chang Chang was surprised: "It's so difficult!"

"That's not all, the team is constantly writing reports on the research results, so that the production system can understand the development work and test the latest methods in real time. In order to achieve a high yield, you have to deal with quality control, and in short, the entire Intel company is a huge information exchange. Dr. Yu said slowly, "And Xiantong, everyone fights independently, drinks tea, chats, and most of the products can't be put on the production line, so there is very little interaction." ”

Chang smiled and said: "I've been told by you that I'm talking nonsense, is it drinking tea and chatting or interacting and collaborating?"

Dr. Yu said: "Your current role is the chief designer, just guide the direction, and you don't have any manufacturing experience, design is your strength, you can only do the low-level design things, as for the rest of the things, after the production line is completed, others will solve it." ”

After thinking for a long time, "I understand. ”

With this discussion, I also made up for the knowledge of management for a long time, and began the development of 32-bit microprocessors with a group of students.

The development goal has been set for a long time, the number of transistors is controlled at 30,000~40,000, the performance is at least 4 times faster than the 68000 processor that appears at this stage, and the addressing space must reach 64 trillion (for the sake of cost, otherwise it is 4GB).

At the same time, the time was estimated for a long time, and the development cycle was set at about 10 months.

Dr. Yu looked at the long-term development plan and was surprised: "How can it be so short! ”

Chang said: "I think it's okay, it seems that there is still a surplus." ”

Dr. Yu said: "You are too crazy, Intel used hundreds of people, and it took a few years to release the 16-bit processor, and now the 32-bit processor has not come out, you don't have any foundation and want to do 32-bit directly, the time is still so tight, how can you achieve it? And the 32-bit processor only has 30,000 transistors, how to do this?"

"There are two reasons!" said Chang Zhizhi on his fingers, "First, don't compare my stuff with Intel's, they are traditional complex instruction processors, and they also use microcode control, just designing this control part has to be exhausting, and it is not easy to find the wrong place." Mine is different, due to the few instructions, the control part uses hard wiring, the design is simple, easy to implement, and easy to manufacture. ”

"Of the 30,000 transistors, only 3,000 transistors are needed for the control part, and the rest of the parts are regular and identical, and can be fully automated. Second, I've got a lot of ideas about designing this processor, it's just a matter of implementation. Based on the above, I think 10 months is more than enough. ”

Dr. Yu read the plan again: "Sounds reasonable, streamlined...... If it's true as you said, the factory has not yet been built, and your things will be shipped!"

Chang smiled and said, "Yes, you can't let me wait for the rice to be cooked." ”

Dr. Yu said, "Then let's wait and see who finishes it first." Although you have your reasons, but I reserve the opinion, 10 months, 30,000 transistors, this is also outrageous. ”

"Bet!" stretched out his hand for a long time, "whoever loses will eat lobster." ”

"Okay!" Dr. Yu agreed without hesitation.

snickered for a long time, this 32-bit chip has been brewing in his heart for a long time, it is estimated that this will be Berkeley's Patterson and others are working hard to develop their RISC-I, this thing should also come out in 82, and he has to hurry.