Japan has almost succeeded

Editor’s note: This article comes from the WeChat public account “the world of magic iron” (ID: jiangpeiyu0916), author Zheng Zhuoran. This article was first published in Tencent.

Key focus:

1. If the idea of ​​Sakamura is successful, the entire IT cake of Bluestar will become a Chinese meal for Japanese companies.

2, Japanese companies are also very wronged, feel that they have been bridged by Intel. They manufacture and sell 8086 and 8088 chips for Intel. After working hard to help X86 architecture become the “world standard”, Intel has cut off the Japanese company. The 32-bit CPU is licensed to enjoy the fruits of the monopoly market alone.

3. Having your own original IP core design capability is of course a top priority, because the purchase of an IP core license may be stuck at any time, but more important than this is the independence and integrity of a country’s sovereignty.

More than 30 years ago, it was the time of the CPU and the operating system, and the market of the squadrons was about to usher in a unified situation. At this time, Japanese companies broke their ambitions in the operating system and processor (CPU) field, intending to stage the legend of the DRAM market to crush Silicon Valley. However, Japanese companies were forced to shrink their pressures and slash their swords from the palace, which made Intel and Microsoft smoothly monopolize the desktop PC field.

This is a forgotten history.


Microsoft founder Bill Gates is so eloquent to people inside and outside the IT circle that Professor Nakamura of the University of Tokyo in Japan is only known to insiders.

坂村健 is a member of the Japanese government’s IT think tank. He is best known for his idea of ​​“computing everywhere” at the age of 29 (1980). The core is to build Japan’s own operating system and CPU. Open strategy to build an ecosystem.

The idea of ​​Yu Cunjian may seem quite normal today, but it was very advanced in 1980. At that time, the CPU was only a sideline of Intel’s not making money. The company’s main business is DRAM (random dynamic memory, commonly known as memory); Microsoft has just been established for 5 years, and later the famous “Wintel” alliance has not yet seen a word, IBM compatible machine It will take another three years to come out.

A history of forgotten: Japan and the United States compete for the dominance of the CPU operating system


At that time, it was still the era of computer development. Today’s giants such as Apple, Intel, Microsoft, etc., are still small enterprises, industry alliances, software and hardware ecology, open source code, etc., which can influence the direction of the industry. Silicon Valley is still a strange existence, so it is an amazing creation that Sakamura can come up with.

And 1980 was also the year when the Japanese economy was opened. Many industries such as automobiles, steel, shipbuilding, home appliances, and semiconductors, from the backwardness of the United States to the one-off industry in the United States, became the main sales force in the US market and successfully staged. The strong dragon hard pressed the head of the snake” drama.

In this case, Nakamura proposed to establish a computer architecture different from Silicon Valley. It can be said that Japan was the natural growth of technology giants at that time, and it led the Silicon Valley from innovation.

If the idea of ​​Sakamura is successful, then the leading IT industry will develop in Japan. Japan will not only dominate DRAM, the core component of the computer, but will also dominate the CPU and operating system that will arise later. The entire Blue Star IT cake will become a Japanese dish.

Japan’s ambitions in the CPU and operating system markets

In 1984, the TRON project led by Yu Cunjian was listed.

The TRON project has become a mature and mature face. It not only develops new operating systems, but also considers the optimal relationship between the CPU and the operating system. Therefore, it proposes to develop the TRON VLSI CPU architecture.

TRON VLSI CPU architecture is a new CPU architecture. It is aimed at Intel’s philosophy of CPU compatibility. It believes that compatibility bears too much historical burden and slows down the CPU. In other words, Japanese scientists such as Sakamura want to start the TRON VLSI CPU architecture from a blank sheet of paper, writing DRAM legends like DRAM.

Emphasizing performance, abandoning compatibility, this concept was later used by Power processors, which is more powerful than Intel’s X86 architecture processor, but it is ecologically difficult due to incompatibility, even with Motorola and IBM. After the support of Daxie, it was finally lost to Intel who took the compatible route.

The situation of Silicon Valley Technology’s infighting has not been staged in the TRON project. On the one hand, Japan has a tradition of close cooperation between government and industry, and the TRON project is also a path of official production and learning; on the other hand, large Japanese companies have seen another gold mine outside the DRAM from the TRON project.

And the Japanese have the ability to dig into this gold mine. Previously, due to the reduction of CPU manufacturing costs by Intel and Motorola, the 6800 and 8086 processors were outsourced to Hitachi, Mitsubishi and NEC. Among them, Intel is more radical, to solve the problem of tight cash flow, and to meet the requirements of more than two suppliers, such as IBM,Founder and CEO Robert Noyce sold the technology license for the 8086 processor to Japanese companies such as NEC.

As a result, Japanese companies such as NEC, Hitachi, and Fujitsu not only have the manufacturing capabilities of the CPU, but also the 8086 processor. They can modify the 8086 processor to ensure compliance with the TRON VLSI CPU architecture specification. Therefore, the TRON project. Welcome. In fact, the first chairman of the TRON project was the then president of Fujitsu.

After the Japanese company joined the TRON project, in order to ensure that the 8086 processor supports the TRON open operating system, immediately contact their overseas companies, concentrate resources on writing a compiler based on the TRON VLSI CPU architecture, and prepare the corresponding software development platform. .

It seems that the Japanese will almost succeed, and will continue to write legends in the CPU and operating system areas after the DRAM market.

Andy waved a sickle to cut Japanese leeks

The Japanese did not expect that the success in the DRAM market would end their dream of ruling CPUs and operating systems.

In 1985, when the 386 processor was about to go on the market, Intel CEO Andy Grove was determined to wave the stick and keep control of the X86 architecture processor firmly in his hands. At the time, not counting the 80286 processor, Intel only issued licenses for the 8086 and 8088 processors, so there are many masks for the two processors in Silicon Valley and Japan. The mature process of making these two processors.

Statistics show that the total shipments of the 8086 and 8088 8-bit processors in 1984 reached 75 million, which is 7.5 times that of the 80286. This means that with the popularity of these two processors, the ecosystem of the X86 architecture has been established. With only the 8086 processor, software companies around the world have invested billions of dollars to write programs for it. .

Now, it is a good time to cut the leek.

There are two important factors that prompted Grove to slash a Japanese leeks. He hated Japan for killing American companies in the DRAM market. During a 125% plan, at a sales meeting, Grove said: “Intel is the last hope for the US electronics industry to face the Japanese electronics industry.” The final outcome is that Intel has quit the DRAM business and confessed to the Japanese.

Now, in the field of CPU, Intel finally ushered in the moment of revenge. More importantly, Grove is also holding the same big killer “Semiconductor Protection Act.”

The introduction of this law is also the result of Grove’s strong push. As early as 1979, Grove believed that Intel needed a more powerful way to protect its intellectual property. In 1980, he asked the company’s legal counsel to lead lawmakers to visit the company’s chip design department, showing the lawmakers,Designing a chip is a tough job, protecting the copyright of the chip (including the mask and instruction set) to prevent foreigners from copying, which is beneficial to the American industry.

After four years of promotion, the Semiconductor Protection Act of 1984 was passed. A legal counsel from Intel recalled that “at least 1/3 to 1/2 of what I drafted in this bill was retained.”

The knives are in hand. Intel began to accuse NEC of quietly improving the 8086 and 8088 chips and selling them in Japan as their own. When selling new CPUs, NEC refused to pay Intel any royalties. Subsequently, Intel filed a complaint and sent NEC to court. The focus of the litigation debate between the two parties was on the microcode (instruction set). As a result of the court’s final judgment, Intel owns the microcode copyright of X86, but Intel cannot claim from NEC because of its own license management issues.

Despite the failure to make a successful claim to NEC, Grove got what he wanted: gained exclusive control over the X86 architecture.

Japanese companies are also very wronged. They feel that they have been bridged by Intel. They manufacture and sell 8086 and 8088 chips for Intel. After working hard to help X86 architecture become the “world standard”, Intel has cut off 32-bit Japanese companies. The authorization of the CPU to enjoy the fruits of the monopoly market alone.

On the other side, Motorola also launched a lawsuit against Hitachi and reclaimed the authorization of the CPU.

The road to Japan’s purchase of US CPU licenses has been completely blocked.

Go to the embedded world

After the development of the TRON project was blocked in the CPU field, Japanese companies pin their hopes on the operating system, but this dream was also ruined by 1989.

After the United States learned that the Japanese government had to install the TRON system into the school’s computer, it pulled down its face and threatened to list TRON as one of the unfair trade barriers. At this time, Japan-US semiconductor trade friction is right, the United States hit the Japanese DRAM industry on a global scale, making many Japanese companies fall into a psychological shadow. Many Japanese computer companies are worried about losing the US market and interrupting the connection with TRON. Muramura said that he was extremely disappointed.

To avoid angering the United States, the Japanese TRON project was forced to scale down, modify the vision, and let the “ubiquitous computing” remove the PC option. The TRON VLSI CPU architecture also turned to MCUs that would not compete with the US (micro control unit) , that is, a single-chip computer), to the embedded field.

In the field of DRAM, Japan was able to shine in the 1970s and early 1980s, mainly because the US government did not realize the strategic position of the chip. After the mid-1980s, with the introduction of the Semiconductor Protection Act of 1984, the IT industry, including chips, was established as a strategic emerging industry in the United States.Industries that need to use national security excuses for protection. Japan is trying to challenge the United States on emerging CPUs and operating systems. Of course, the United States can’t stand it. So when TRON is still in its infancy, it will not give it a chance to re-enact the DRAM market to break down Silicon Valley companies.

In the eyes of the American industry, with the current development of TRON in the embedded field, the big stick is just right.

TRON (system + MCU) moved to the embedded field, with the concept of open source, and the extensive participation of Japanese manufacturing, became the brain of electronics, automotive, industrial equipment, as early as 2004, the TRON system was installed Of the 3 to 4 billion home electronic devices, more than 150 million installed units of Microsoft Windows, TRON architecture MCU shipments also greatly exceed Intel CPU.

A forgotten history: Japan and the United States compete for the dominance of the CPU operating system

Toshiba Fujitsu’s MCU products, which serve as brains for home appliances, automobiles, and industrial equipment, have extremely low power consumption.

Because of the influence of TRON, now the rivals Microsoft and Intel have to join. Someone has posted the bill, even if he is charged a 10 cent license fee to each TRON user, he has the strength and Bill Gates to slap on the table of the world’s richest man.

Miss the best time window

The problem is, without the CPU license of Silicon Valley companies such as Intel, can Japanese companies design CPUs independently? The answer is possible, because in the mid-1980s it was an excellent time window for Japan to independently develop CPUs.

At that time, Intel had just regained control of the X86 architecture, and the CPU war was not extinguished. Later, CPU battles between Intel and the AIM Alliance (Apple, IBM, and Motorola) broke out. Later, ARM released the RISC instruction set and also competed with the Intel X86 CISC instruction set.

In other words, the CPU market is not fixed, everything is possible.

More importantly, the CPU structure at that time was far from the complexity of today. The number of transistors in the 286 processor is only 100,000, the number of transistors in the 386 is about 400,000, and the number of 486 transistors is about 10 million. As long as the CPU package is opened, the circuit design of the CPU can be obtained by polishing the chip layer by layer. Silicon Valley companies often accuse Japanese companies of doing this.

A forgotten history: Japan and the United States compete for the dominance of the CPU operating system

The number of transistors in Intel’s early CPU

The Japanese company has many years of experience in CPU foundry and magic transformation, and has the ability to design and manufacture its own CPU. In fact, after being cut off by American companies, Toshiba, Fujitsu, and Matsushita have separately designed and manufactured the CPU of the MIPS instruction set, but they are doing brains for home appliances, automobiles, etc. If the performance is improved, it is a replica of the ARM processor.

But because of the fear of offending the American industry, Japanese companies dare not go one step further, designing and manufacturing high-performance CPUs based on MCUs, and obediently cutting the TRON operating system and actively moving away from the PC field.

Japanese companies have shrunk their CPUs and operating systems, and finally made Linux and ARM processors.

In terms of self-developed CPU, having its own original IP core design capability is of course a top priority, because the purchase of IP core authorization may be stuck at any time, but more important than this is the independence and integrity of a country’s sovereignty. Otherwise, no matter how good the time window will be missed.