Chapter 712 of the main text: the fusion of CISC and RISC

4823737.html

Next chapter

Lin Qiangsheng led John Kleber and Zhang Rujing to visit most of the East Star Semiconductor Company, which has 12 buildings of different styles, as well as two wafer factories, a chip processing plant and a packaging and testing factory.

And now Johnson Lin's public relations team is still negotiating with the San Jose County government, lobbying senators in the city to build Cisco Systems headquarters for a new piece of land, which will be another pillar of Johnson Lin's group!

John Kleber may not be very familiar with Lin Johnson, he only knows that he led the creation of IBM's reduced instruction set research and development project in 80 years, and he is one of the founders of the early development of Japan.

Zhang Rujing is very famous in China, and he will be the founder of SMIC, a wafer manufacturer founded in Shanghai and one of the most technologically advanced integrated circuit chip manufacturers in China. He has managed more than 10 wafer fabs for Texas Instruments in the United States, Singapore, Taiwan, and China, and after founding his own company, Zhang Rujing brought the perfect management technology of Texas Instruments to the factory, making SMIC a pivotal semiconductor company in China.

And now Zhang Rujing left Deyi more than ten years early, but Lin Qiangsheng still believes in his management talents.

The four of them walked into the semiconductor real center, Lin Qiangsheng looked back at Ibeka who was looking around, and said, "You take care of yourself, don't lose it!" ”

Ibeka smiled and said, "Don't worry about me, I'll follow you closely!" ”

John Kleber and Zhang Rujing didn't know what was going on with this woman, but they guessed that this might be Lin Qiangsheng's bodyguard or something, so they didn't care about her.

Lin Qiangsheng said to John Kleber: "Mr. Kleber, I have always had a doubt! ”

Johan Kleber paid attention to his speech and demeanor, and said humbly, "Sir." You can call me Yohan, everyone calls me that at IBM, what do you have any doubts? ”

Lin Qiangsheng smiled helplessly. Four or five out of ten Americans named John, and he nodded and said, "John." That's it, I've been wondering a question! Can't complex instruction sets and reduced instruction sets be used in parallel? For example, can an 86 chip from Intel be integrated with a reduced instruction set processor? This allows computers to take advantage of the reduced instruction set processor when operating some single task. We know that the daily S power consumption is low and the speed is fast, which can significantly improve the processing speed of the computer, so that the processing power of P can be increased several times is not a problem! ”

Lin Qiangsheng's words suddenly stunned John Kleber, who was proficient in Japanese s.

In the early years of the development of the Japanese S system, some great gods who advocated the reduction of the instruction set always felt that the Japanese S was the future development direction of the computer, and they completely abandoned the complex instruction set of IS. The belief that the day S will defeat IS sooner or later is completely incompatible. On the other hand, the gods of the IS God Cult also have similar thoughts, thinking that the day is just a game for a bunch of misguides, and they don't take the reduced instruction set in their eyes at all!

So hearing Lin Qiangsheng's statement, John Kleber, who reacted, said in shock: "You mean, integrate the Japanese S processor into the IS processor?" ”

He had never considered this idea, how could complex instruction sets and reduced instruction sets be compatible!

But

Lin Qiangsheng nodded to Kleber who was thinking and said: "Take the 86 architecture of the IS complex instruction set as an example, you can learn from the idea of the Japanese S reduced instruction set, such as integrating some microinstructions into 86!" ”

He looked at the puzzled Kleber and said, "Even." In the future, our 32-bit Japanese S processor can also integrate a 16-bit complex instruction set, so that the efficiency of the processor will be higher, and the two can actually learn from each other, right?! ”

Heard Lin Qiangsheng's words. The high wall between the is and the sun in John Kleber's heart collapsed in an instant, if the boss just mentioned it just now, then now it is already a thorough way to teach himself how to do it.

He immediately calculated in his mind that it was feasible, and the result was completely feasible and correct!

But Kleber was still bound by his previous thinking, and he said with some hesitation: "But in this way, the classic complex instruction set of 86 will become a freak that is not complicated and not streamlined, can this Intel do it?!" ”

Lin Qiangsheng smiled. Faced with Kleber's somewhat pedantic question, he nodded and said, "Intel definitely won't do it." But we don't still produce 86 processors, and your old employer, IBM, is doing the same! ”

"Ugh," Kleber had to nod in acknowledgement. Although IBM and Intel have a deep cooperative relationship, at the beginning of Intel in the IBM chip factory to order chips, IBM also took the opportunity to clone Intel chips and made its own 286 processor.

Lin Qiangsheng continued: "Just like that, we can also use the dual cores of the Japanese S processor, and the two cores are designed to use the same interface and bus!" Even we can apply this 'dual-core' approach to the 86 architecture, and two 386 chips can be used together, so that the performance of the computer can be greatly improved! ”

This is the concept that prevailed in the IT world in the previous life, if Lin Johnson didn't know, it would be too autote, it must be very interesting to put forward such a concept in the mid-80s, and John Kleber, as a senior engineer and former technology executive at IBM, also knew the novelty and feasibility of these concepts, he was just stunned by Lin Johnson's bold idea for a while.

At this time, Zhang Rujing next to him said: "Mr. Kleber, I think the idea proposed by Mr. Lin is very good, and it is very likely to be realized!" ”

Zhang Rujing said to the two people who turned their heads: "Let's not mention the integration of the reduced instruction set into 86, the concept of dual-core mentioned by Mr. Lin is very feasible!" It's the same concept as a large computer with many chips, except that it's an integration of countless chips, and a dual-core has only two cores! ”

In fact, Zhang Rujing is more in favor of Lin Qiangsheng's idea of integrating a reduced instruction set in the 86 architecture. If you want to achieve a dual-core computer, it will be much more difficult to achieve, and the combination of two cores will compete for front-side bus bandwidth, which will become a bottleneck in system performance, and if you want to solve this problem, you need to redesign the PU, which is no different from re-developing a processor, which is much more difficult, and the integration of the reduced instruction set is much simpler and feasible, but the former is better understood.

Then, Zhang Rujing said to Lin Qiangsheng in approval: "Mr. Lin's proposal to integrate the reduced instruction set on 86 cleverly crosses the gap between IS and Japanese S, combining the two together, although blurring the boundaries between them, but this middle-of-the-road approach can indeed significantly improve the performance of PU, which is almost predictable!" ”

"Isn't it, Mr. Kleber?" Zhang Rujing smiled and asked rhetorically to Yohan Kleber.

At this time, John Kleber had to nod heavily, he looked at Lin Qiangsheng and Zhang Rujing, one is the future boss, the other is the leader in the same industry, so he said: "Since you all think it is feasible, then Mr. Lin, I ask for an increase in the budget, we can add a parallel project in addition to the development day s processor, set up a project team, and integrate the reduced instruction set into the 86 architecture!" ”

At the beginning, John Kleber reported a budget of $12 million to Johnson Lin to develop a processor with a thin architecture, and now Kleber has taken over two tasks, Johnson Lin is very happy, he said: "Yes, then I will approve your budget of $4 million!" ”

John Kleber made a gesture and smiled, "OK, no problem!" ”

Lin Qiangsheng and others entered the real yàn room and made a brief introduction to Zhang Xi and other R&D personnel in the real yàn room, so everyone knew that Kleber and Zhang Rujing, one from IBM and the other from Texas Instruments, are both famous companies in the semiconductor industry, a blue giant, and an old-timer in the industry!

After that, John Kleber stayed in the real room, and Lin Qiangsheng and Zhang Rujing left the building and got on the battery car to come to the wafer factory of Dongxing Semiconductor.

The driver was a young man from China, he drove the battery car directly into the factory, and after changing into dust-free clothes, Lin Qiangsheng, Zhang Rujing and Ibeka walked into the wafer production workshop.

Lin Qiangsheng said to Zhang Rujing: "In the United States, the replacement of wafer size is generally about ten years, and now the world is the world of wafers, and Intel has mastered the manufacturing technology of 6-inch wafers, and their current wafer factories are mainly 5-inch. Now our wafer fab is also gradually replacing the manufacturing equipment of 5-inch wafers, this is the upgraded one, and the other will be upgraded before the end of the year! ”

Zhang Rujing looked at the equipment in the workshop, and was no stranger to this kind of advanced automatic and high-precision things, he said: "Intel's technology is indeed developing rapidly, from 286 to 386 fully proves the correctness of Moore's Law!" ”

Lin Qiangsheng said helplessly: "Intel has money, and it can constantly update the manufacturing process and improve the architecture." Every two years, they launch processors with smaller core areas and more advanced processes, and every two years they launch processors with newer architectures. In this regard, we and other processor manufacturers cannot compare with Intel! ”

Intel's development model, in fact, is to fight for money, with money there are talents, a large amount of money is constantly invested in technology upgrading and research and development, in order to make Intel's chip technology continue to improve, this is actually a kind of research and development competition.

The Americans are good at engaging in military competition, but in essence it is a fierce competition between R&D fields!

This is very strange to the Chinese, China has always been the one who is catching up, and has never known what it is like to be ahead, let alone compete with the world!

Lin Qiangsheng can't do this now, both in the United States and China, there is a great deal of uncertainty between China and the United States, which is bothering him. Besides, even if he wants to invest a lot of money, there is no talent base in China, and now many domestic elites in the computer industry are in the growth stage, and it is too late to make up for the knowledge gap caused by these years! (To be continued.) )