After a visit of more than a month, the "Ten-Year Plan" team gradually gained a certain understanding of the technological level of the domestic semiconductor industry.

A simple route has been made for the development of the domestic semiconductor industry: 1. Production and R&D should be combined to produce products that meet the market demand. 2. Follow the law of Moore's Law, and upgrade the process every 18 generations. 3. Encourage and develop large-size wafer fabs, strive to popularize 3-inch and 4-inch wafer fabs in 85 years, and 6-inch wafer fabs will become the mainstream within the ten-year plan. 4. In terms of semiconductor technology, 5-micron technology was popularized and promoted (it became the mainstream in 84), 3-micron technology was developed (promoted in 86), and 1-micron technology was tackled (it entered the market in 88).

This goal is naturally much higher than the historical goal, after all, after the historical goal is achieved, the gap with international counterparts has widened. However, this goal is to narrow the gap with international manufacturers.

Semiconductors are mainly used to process wafers, in which the size of the wafer is naturally as large as possible, and the larger the wafer radius, the more transistors can be cut. At this stage, silicon wafers are the most basic raw materials in the semiconductor industry.

For example, Jiangnan Radio 742 Factory in Wuxi, Jiangsu Province, introduced an IC production line from Toshiba of Japan, which includes a 3-inch wafer factory and a complete set of processing and packaging production technology, with a total investment of 66 million US dollars, and a monthly production of 10,000 3-inch silicon wafers and an annual production of 26 million IC chip products. However, this production line can only be used for the production of IC chips for TVs, although the production capacity is not low, however, it cannot be repurposed for the production of other chips.

Most of the other domestic wafer factories only have 1.5-inch and 2-inch wafer production capacity. The size of the wafer is too small, and the raw materials supplied to the chip processing industry are naturally insufficient, so that high-quality wafers have to be imported from abroad on a large scale.

The process of processing wafers is naturally as small as possible, and the more transistors can be cut under the same wafer area.

Tsinghua University in China in the 70s manufactured the domestic architecture of the WeChat computer DJS050, that is, because the semiconductor process is too backward, the CPU chip used more than 30 integrated circuits to piece together, to describe it a little awesomely, our first generation of domestic DJS050 personal computer in the 70s is more than 30 core chips, but ...... The number of transistors of 30-core chips is not as good as other single-core CPUs, and multi-core is a helpless move, which can integrate more transistors into single-core chips, and the higher the performance of the main frequency, the better. In the case of no choice, the single core cannot improve more performance in the short term, so it is forced to engage in dual-core, quad-core, and octa-core ......

Because, DJS050, a multi-core computer with more than 30 cores, is not as good as an 8-bit Apple computer, and it is not practical, so after the research is out, it will not be mass-produced.

Most of the domestic technical level stays below the 10-micron process. In a few factories introduced in recent years, the process is about 8 microns.

Lishan Semiconductor's 5-micron process is the most luxurious equipment that the aerospace military system can make up, and the 5-micron technology in China is already the strongest in industrial mass production. Moreover, the equipment has not reached the level of complete localization, and the 5-micron lithography machine is still imported from Japan.

The "Ten-Year Plan" group does not want domestic semiconductor factories to rush to make the most technically complex CPUs. On the contrary, CPU manufacturing companies are limited to 10 to avoid overly dispersed resources, and each one is not big or strong.

However, the DARM memory with a low technical threshold is to encourage domestic semiconductor factories to vigorously do it, which has a large market demand and low requirements for the technical threshold.

For example, in the development of DARM memory capacity, there was already 1K DRAM in 70, 4K (10 microns) in 1974, 16K (5 microns) in 1976, 64K (5 microns) in 1979, 256K (1.5 microns) in 1983, and 1MB (1 micron) in 1986......

According to the advocacy of the ten-year planning group, coupled with the solid orders of Xinchuang Electronics, from 84 years onwards, the annual purchase of DARM from China is not less than 1 billion yuan. With the continuous release of domestic memory production capacity, production and sales have also been blowing out.

By the end of the 80s, after the United States gradually withdrew from the memory market, the three major powers in the world's memory industry were China, Japan and South Korea. In this, Japan defeated the memory of the United States with cost performance, and China and South Korea took most of the low-end cake from Japanese manufacturers at low prices in the memory market, and China and South Korea fought the most fiercely. The final pattern is mainly that Chinese manufacturers almost monopolize the low-end memory market, South Korea is the hegemon of the mid-range memory market, and Japan still occupies a leading position in some high-performance memory markets with its technological level.

But this market is not static, and the final result of competition in the memory market is to encroach on the high-end from the low-end, because the low-end market demand is the largest and the production capacity is also the largest. Without capacity, technology will not advance too quickly. As much production increases, all kinds of subtle discoveries and feedback in the process may eventually be transformed into technological advantages.

"Although the domestic semiconductor technology is relatively backward, the low-end products are promising!" Yu Youcheng said with a smile, "The DRAM market is very broad!"

"That's right! Although the 4K memory is a little low-end, there is still a certain market for handheld computers and low-end computers!" Ni Guangnan said with a smile, "16-bit computers are popular, and they can't afford to use them in China, but the Great Wall Computer Company is engaged in mass production of 8-bit Pangu ......."

"8 Pangu?" Yu Youcheng smiled and said, "Pangu does not have 8 digits, the minimum is 16 bits, but they are abducted and deceived under the brand of our company." If it weren't for the Great Wall Company, I would have been ready to complain about them. ”

The entrepreneurial work of the Great Wall Computer Company was originally supposed to be a PC compatible machine, but because the XRM architecture chip is now more popular in China, the price is also cheaper, the market price is 30 yuan, you can buy an X8 chip, although it is an 8-bit chip, but it is not inferior to the 16-bit 8086, 8088 chip, but, the technical potential has been tapped, even if the process level is improved, there is not much room for technical upgrades.

Therefore, after the technology of the 8-bit chip was developed, it was abandoned by the new venture electronics company and authorized to Lishan Semiconductor Co., Ltd. for production, and only charged 1 yuan for each chip. Even though the copyright fee is so low, due to the strong demand for handheld computers, Lishan Semiconductor has to produce more than 60 million X8 chips a year.

These chips are mainly supplied to the palm entertainment platform and used as CPU chips for handheld computers. However, some domestic factories also purchase this chip to develop industrial computers.

Of course, the first generation of domestic Galaxy supercomputers also used this cheap 8-bit CPU as a chip for low prices, so they piled up a supercomputer with a peak of nearly 200 million times, and its performance was stronger than the 100 million peak computers in history.

In addition, at present, domestic small computers are also gradually adopting this cheap and beautiful chip.

With the increase in the number of users, the ecosystem of application software is gradually improving, not like at the beginning, the entire platform is almost blank. If you want an operating system, you have to write it by looking at the assembly, and if you want to implement a certain function, you have to write the code yourself. Early adopters are almost hard pressed and can't be forced anymore. Of course, these hard work is also rewarding. Many programmers originally used their own code, but found that others needed it, so they began to sell the programs they wrote.

This bears a striking resemblance to what is happening in Silicon Valley. Of course. It's just that limited to the scale of computing users in China, many software authors are equivalent to self-employed, making a very small number of copies, and only when the old copies are sold, will they engrave a few new copies.

Now the best computer sold in China, of course, is the Pangu host, which can make most of the small computers withdraw from the stage of history. The Pangu host, which is priced at only 3,000 yuan in the domestic market, is not inferior to a small machine with a performance of more than 100,000 yuan. And...... Users of the Pangu console also have a secret secret, this platform has a lot of game thieves, and every now and then there are a bunch of fun new game updates. Even some customers who are busy for work and scientific research have time to play new games occasionally. If it passes early, people are queuing up to use the small computer, this month to apply for the next month before your turn, and, at most, you will use a few days, a few days must be given to other units to use, and the daily cost is thousands of dollars, in this case, those precious little machines are basically a race against time, precious time can not be wasted.

The Pangu host is about three days of rent compared to the minicomputer, in this case, it is difficult for the Pangu host not to be popularized.

As for the 8-bit "Pangu compatible machine" produced by the Great Wall Computer Company, in fact, many programs that can only run on 16-bit computers are not compatible at all. Only a very small number of Pangu platform programs can be successfully run on the Great Wall compatible machine.

However, many of the 8-bit programs on the Great Wall compatible machine can be successfully run on the Pangu platform, after all, the architecture is XRM, and XRM16 is designed to be compatible with XRM8.

But the 8-bit Great Wall computer with very poor performance also has a huge market, because - cheap!

The price is so low that it costs only 498 yuan without the monitor. As for the monitor, you can use the TV set in your home instead.

In this case, people who can't afford to buy Pangu computers and PCs can replace them with Great Wall computers first.

Of course, for Great Wall Computer, the production of 8-bit computers is just a stopgap measure. Later, it seized the general trend of the gradual opening of technical standards of Pangu computers, and began to respond to the call of Pangu computers in 85 years.

Even in the 90s, Pangu further opened up its technical standards, allowing more software and hardware manufacturers to produce the hardware of the Pangu platform. However, Great Wall's hard disk has been strong, accounting for 50% of the market share of Pangu computer camp's hard disk.

Later, when the global mechanical hard disk market was reshuffled, Great Wall was even ready to acquire Western Digital to challenge the position of the hegemon in the hard disk market. Of course, it was later obstructed by competitors such as Seagate, and the United States was unwilling to let Chinese manufacturers become the champion of the mechanical hard disk market, so Great Wall could only continue to do four and three, and continue to wait for the top three Seagate, Western Digital and Toshiba to die, and gradually let it take the position.

Of course, the Great Wall hard disk in this world, not as regrettably collapsed as the future history of the parallel world, and the biggest contributor is not the Great Wall Computer Company. In 85 years, the new venture electronics company applied for a heavyweight technical patent - MR (Magneto Resistive) technology, this new type of head adopts the method of magnetic induction writing and magnetoresistive reading, which greatly improves the sensitivity of the head and greatly improves the work efficiency of the hard disk, and at the same time, the storage density of the platter has increased dozens of times compared with the previous 20Mbpi (bit / per square inch), laying the foundation for the huge increase in the capacity of the hard disk.

Relying on this technology, domestic hard disk manufacturers counterattacked, and launched a cheap 100MB capacity hard disk in 85 years, which suddenly made the mechanical hard disk of the 20MB~50MB capacity personal computer market on the market collapsed.

Later, hard disk manufacturers in Japan and the United States also obtained the right to use the technology at the cost of paying a patent license fee of $5 per hard drive. However, domestic manufacturers are the first to seize a large part of the market with a time difference. For example, Great Wall's hard disk not only has a market in the Pangu camp, but also has always been a top 10 hard disk brand in the PC camp.

……

"Popularize and promote 5-micron technology, develop 3-micron technology, and carry out 1-micron technical research. "This was originally a goal set during the Seventh Five-Year Plan in '86, but now this goal is being proposed.

Because that is to say, the semiconductor process development plan of the Seventh Five-Year Plan era, when this goal was formulated, it was already backward than the technology that has been popularized in the semiconductor industry, and after the completion of technical research, compared with the world's counterparts, not only did not narrow the technology generation gap, but made the domestic process further backward.

The bigger problem is that this technical research mainly stays in the laboratory, and after the 1-micron process is realized, it is only used to provide laboratory chips, and does not have the ability of large-scale industrial production.

Even if this equipment is made, it can only be announced to foreigners that I have achieved this process, but ...... Large foreign companies sent people to visit, and sure enough, the Chinese laboratory has already conquered the 1 micron process. The 1-micron process production equipment, which was originally blocked by technology from China, can already be sold to China.

To put it simply, foreign manufacturers have made profits by dumping technical equipment sold to China on the one hand, and on the other hand, they have suppressed the development of China's own semiconductor equipment manufacturing industry.

The domestic semiconductor industry imports equipment from overseas, and the domestic equipment industry naturally cannot get orders, so it cannot develop.

Even if you have money in the future, if you want to develop, because the market and technical standards are monopolized by others, and the technical threshold is too high, it is becoming more and more difficult to get through.

Now we are developing materials, equipment, and other foundations from scratch, and once we have the results, we will apply them to industrial production on a large scale. Even if the production process is inferior to foreign products, it is okay to encounter various problems, and even if the development problem is solved, this is a necessary stage of development.

Including the United States, West Germany, Japan, the Netherlands and other semiconductor equipment manufacturing powers, their equipment is not perfect, but there are a bunch of defective products and catastrophic losses in the process, after that, because the problem is found and solved, after that, technology and experience have been improved.

If you dare not use your own technology and equipment because you are afraid of technological gaps and quality problems, then advanced technology and production capacity will never fall from the sky.