The establishment of the new venture semiconductor company was largely driven by Dr. Yu Youcheng. Although, the current team of the new entrepreneurial chip design and process can be called full of talents.

However, more than ten years ago, without Yu Youcheng's experience in supervising chip R&D and process manufacturing from Fairchild and Intel to the new venture electronics group, Lin Qi's chip dream may have to be delayed for many years to be realized.

In the chip manufacturing industry, the most difficult thing is actually at the beginning. In the beginning, it was impossible to do a project without experience, even if you read professional books in the semiconductor and chip manufacturing industries, and without actual experience in R&D and production. Therefore, the fastest growth at the beginning should be from the practice, so that the newcomers are exposed to more practices in the production and design process, with the accumulation of time, book knowledge and work experience are integrated, some newcomers become old birds, and even, will do continuous innovation in the actual work. The innovation of one or two people is actually very limited, and it is impossible to play a leading role in more and more sophisticated design, and more importantly, more people continue to contribute, and technology can continue to develop.

At present, new start-up semiconductor companies are able to self-cultivate and upgrade themselves. After the newcomer comes in, after a few years of training, he may become a senior talent in the industry, and he can have some say in the field of design and production.

In a sense, the top scientists no longer grasp the specific design work, but more in terms of direction. Or rather, to solve more non-technical problems for the team.

The team that has grown up doesn't even need Dr. Yu Youcheng anymore. Because his experience and knowledge in the early days were very important, but now many innovations in the XRM architecture are no longer led by him, but are constantly promoted and developed by more latecomers.

Therefore, in 91 years, Dr. Yu Youcheng has actually taken a back seat, no longer holding actual positions in Xinchuang Electronics Group and Xinchuang Semiconductor Company, and only retains the honorary title of lifelong chief scientist of Xinchuang Electronics Group, as well as holding shares worth more than 3 billion Hong Kong dollars in Xinchuang Electronics Group. Originally, there was also an opportunity to obtain the equity of the new entrepreneurial version of the semiconductor company, but after making a good return, Dr. Yu Youcheng was relatively bearish in terms of money and equity, after all, the equity of the group held in hand was not less than 300 million Hong Kong dollars per year. It is precisely for this reason that Dr. Yu Youcheng began to hold positions in Tsinghua University and Peking University in Beijing for a higher level of fame, and won the honorary title of academician of the Chinese Academy of Sciences.

After becoming famous, Dr. Yu Youcheng often talked about what happened during his work at Intel Corporation in the United States, but more memories of the development of chips in the new entrepreneurship department and the wafer factory from scratch. To a certain extent, it has also inspired more young people in China who are interested in the semiconductor industry to choose to join the new entrepreneurship department.

As for the current CEO of the new venture semiconductor company, it is the young and excessive Hu Weiwu, who was introduced to the new venture semiconductor company by Dr. Yu Youcheng a few years ago for internship, but now he has become the most important scientist of the company.

Hu Weiwu studied under Academician Xia Peisu and was also a genius in later generations. It is not as dazzling as Huawei's HiSilicon Kirin, but it has mastered the independent and controllable Loongson. Historically, after the advent of Loongson in 2002, although it has been backward in technology, it is more restricted by the domestic process and the poor capital of the Loongson team. Even so, in 2002, Loongson 1, which adopted a reduced instruction set, has achieved 266MHz, and after that, Loongson 2 has achieved 1GMHZ. Loongson 3 is to develop different multi-core versions. Although, since the advent of Loongson, it has faced a lack of market orders, and due to the low output and high cost of a single chip, the Loongson terminal, compared with other chips, does not have any cost-effective advantage. However, Loongson is of great significance, because it is an instruction set architecture patent that is completely in the hands of Chinese Loongson, which can be modified, upgraded and developed independently, without being controlled by others.

has been ridiculed for many years, thinking that Loongson is like Don Quixote, there are already various mature chips on the market, but they have to do chips with independent intellectual property rights. As a result, it has been a dismal operation, and the application of Loongson has been very narrow, so the talents of the Loongson team have led the lowest salary in the industry and the lowest R&D expenses in the industry, but they have been struggling to develop the chip.

But it's a miracle to be able to survive all the time. After staying up until 2015, Loongson became famous because it was purchased by Beidou satellites.

To a certain extent, even Huawei's HiSilicon Kirin chip, although more commercially successful, is still developed and improved on the basis of ARM's patent licensing. Therefore, compared to Loongson, HiSilicon Kirin is more likely to achieve market success. After all, the ARM chip camp, the major chip manufacturers, are based on the authorized technology of the ARM chip architecture, and optimize and improve it themselves. HiSilicon Kirin does not need to consider the issue of market compatibility, but only needs to consider the issue of standing out from the homogeneous products. However, if HiSilicon has been relying on ARM technology licensing, there are still huge hidden dangers. Suppose, ARM suddenly doesn't give permission?

Comparatively speaking, Loongson has more feelings than HiSilicon Kirin.

It's a pity that Loongson was born at the wrong time, if it came out 20 years or even 10 years earlier, maybe the development would be much better.

In general, in terms of talent and level, Hu Weiwu is definitely the top in the domestic semiconductor industry. As for the average business achievement, it is simply because he chose the most difficult path to make money.

If you take the route of easy to stand in front and choose to make ARM-based chips at the level of the Loongson team, it is estimated that there will be no HiSilicon Kirin for a long time.

I am afraid that it is precisely because of this ambition that Hu Weiwu still chose to make domestic chips even though history has changed. It's just that it is no longer Loongson, but in the new start-up semiconductor company to preside over the research and development of chips.

……

Lin Qi went to the dust-free laboratory of a start-up semiconductor company.

Hu Weiwu took a chip and showed it to Lin Qi: "Mr. Lin, this is an ARM architecture chip manufactured by us based on the 0.5 micron process. At present, there is a huge demand for low-power chips for Kirin mobile phones and handheld platforms. Therefore, our 0.5 process production line gives priority to the production of ARM chips. As for the XRM architecture of the Luban 2nd generation chip, it is temporarily produced in advance by Lishan Semiconductor. ”

Although, the leading enterprise of domestic chip architecture, the new process is ready to make ARM architecture chips first, which is a bit wrong. However, Lin Qi smiled dumbly, because, history has changed, ARM is already a member of the new entrepreneurship department, and the patent rights of the ARM architecture are completely owned by the new venture electronics group.

Therefore, the ARM architecture can also be regarded as a domestic chip architecture, although the ARM team is still produced in the UK, but the technology is shared with the domestic XRM team.

In essence, the ARM architecture is very different from the historical ARM, although it is also a reduced instruction set, and it is also low power consumption. However, during the design process, a large number of XRM architecture technology licenses were introduced, and as a result, ARM has begun to become a low-power version of the XRM architecture.

It's just that the compatibility between the two is still a little worse, because the instructions given by Lin Qi are that the ARM architecture must be a low-power chip, and it should not become a low-performance XRM chip because it is considered compatible with the XRM architecture.

You know, if the direction is wrong, it will be more and more difficult to correct it in the future.

For example, Intel's x86 architecture was not designed for low power consumption in the first place. Therefore, even if I realized that low power consumption represented a larger market direction, because the X86 chip architecture was already very complex, it was more difficult to modify X86 than to design a chip with a new architecture from scratch. Therefore, it is difficult to start from the design to reduce the power consumption of the X86 chip, or to reduce the performance and achieve lower power consumption. Either the process is improved, after the process level is improved, the performance is improved slightly, and the power consumption can be reduced.

But no matter how much the power consumption is reduced, it is still difficult for X86 to compare the power consumption with the reduced instruction set chip.

Even other reduced instruction sets, which are successfully commercialized, are difficult to surpass the ARM architecture.

"How much difference is there between the power consumption and performance of the current ARM-based chips and XRM?" Lin Qi asked.

"ARM chips can only achieve 50 MHz at present, and the power consumption of a single core is only 0.7 watts, and the power consumption of dual cores is only 1.5 watts. Hu Weiwu said, "In terms of performance, chips that may not be able to reach the same frequency XRM architecture are at least 30% weaker than XRM architecture chips with the same frequency." However, in terms of power consumption, it is a chip that completely abuses the XRM architecture. The Luban 2nd generation chip, the main frequency is increased to 100 MHz, and the maximum power consumption of a single core is more than 5 watts. The maximum power consumption of the quad-core is more than 20 watts. Therefore, for high-performance desktops and laptops, the XRM architecture is still dominant. As for ARM, it is more suitable for low-power platforms such as handheld computers and learning machines. ”

Subsequently, Hu Weiwu even showed Lin Qi a good thing - a 7-inch handheld!

This is the 3rd generation of Palm Entertainment developed by Palm Entertainment Technology Company, and it has been skipping tickets for many years, mainly because it wants to make trouble. The 7-inch palm entertainment 3rd generation handheld computer, strictly speaking, should be called a small laptop. It's just that on both sides are the keys of the handheld, and after turning it over, it is the notebook keypad.

"The notebook system is also Pangu?" Lin Qi said hesitantly, "What is the practicality of such a small thing? There are problems with heat dissipation, battery life, etc., right? Products that run out of power in one or two hours will not be practical." ”

"Boss, you're overthinking. Hu Weiwu said with a smile, "This platform is a system that has been simplified and modified from the Pangu system, which was originally only compatible with the XRM architecture, and has stripped off a bunch of unused functions and streamlined to the extreme." Modified to the extent that the ARM platform can also be installed. However, it is not fully compatible with Pangu computers, but as long as the software company cooperates with the streamlining and changes, many applications can be ported. After the functions become less, the power consumption of the system itself is also lower, so, this 7-inch handheld computer, theoretically, has a battery life of more than 6 hours. What's even more powerful is that it can not only play games, but also surf the Internet and work. In terms of browsers and word processing software, it is already compatible with the version on the Pangu system. ”