Strong industrial base blessing, China's "core" difficult road

Following this trend, we have seen news of the company's Loongson 3B server in the media today. In fact, this server was released in the third quarter of last year, including the L620-G15, L640-G15, TC4600L three, as well as a desktop office terminal L300-G10, a high performance dual server L620R-S.

Since 2001, Godson has been fifteen years since then, and the young and enthusiastic youth Hu Weiwu has entered the middle age. For many years, the Godson project, which was used by public opinion as a “crook” and “fishing,” has finally seen the blessing of the industrial base. A little dawn.




Godson was originally an experimental project conducted by the Chinese Academy of Sciences. In 2001, there was no Haisi Kirin 920, Rockchip RK3288C. CPU licensing was still high technology in that era. Therefore, the start of Godson started with an IDT chip, using the MIPS instruction set, made Loongson 1.

Then make persistent efforts to quickly make the desktop level of Godson-2, although the technology is still relatively backward, but it is exciting from scratch. Godson is quite lively for a while.

There are listings, investments, and private entrepreneurs are ready to engage in industrialization.

The reality gave Godson a cold water. Loongson only solves the problem. The instruction set uses MIPS and the operating system uses Linux. By 2005, the performance was only that of mainstream computers in 1998. The advantage is that the power consumption is slightly lower, and the cost is not cheap because of the small batch size. The market performance of this kind of thing can be imagined.

Although there are Godson notebooks, Loongson micro-computers, but a performance equivalent to P3, there is no software to use, the price is not cheap how could sell things, so Godson can only rely on the country to raise, research and development of high-tech , but the market can't trade for money.

Without money, less than first-class talents are recruited, the level of products cannot be improved, and even the stability is difficult to guarantee. Some projects in professional fields are not dared to use them after testing. This is a vicious cycle in which Godson sinks into silence.



In fact, China has more than one Loongson chip project. There is a very nimble project that is the god of the Jiangnan Institute. Shenwei is a direct descendant of the legendary Alpha processor.

Alpha was once a legend in the industry. The rise of ARM was a great help for Alpha. Intel's hyperthreading came from Alpha. AMD's entire EV7 bus was moved from Alpha. Alpha led Intel for more than 10 years and proposed a framework for many-core.

However, technologies and markets are not always the same. DEC, which owns Alpha, was acquired by Compaq and then merged by HP. Alpha was displaced to China, and was purchased by Jiangnan as a major asset.

Shenwei processor's floating point capability exceeds that of Intel, and its performance and power consumption ratio is also outstanding, but it is software-dedicated and difficult to popularize. Although several supercomputers were made to meet the needs of some professional fields, they still could not get rid of the state support.

At the other end of the Chinese chip industry, embedded processors, because of market demand stimulation, Rockchip started from MP3, step by step to the tablet, development is very good, also received the favor of Intel, Huawei Hass in order to meet Huawei With its own needs, it has developed a series of CPUs for the Kirin series, and it is also close to world-class standards.

Market-oriented, profitable, R & D combined with market demand can continue to progress, but also can recruit first-rate talent, form a positive cycle, without state support, but also to the country's tax.

The Yangchun Baixue line is a wasteful source of money, and it is in search of a market outlet for good development.

Prism door opportunities

The prism door exposure shocked the world. Although various backdoors have been speculated and sporadic reports, the real exposure has alerted many governments and important agencies.


It turns out that the software and hardware that we often use are so insecure that the information of some important departments will be easily monitored. This gives opportunities for the marketization of Godson and Shenwei.

Spell performance, fight services, fight relationships, international giants advanced technology, ingrained. However, once mandatory security requirements are met, it is equivalent to establishing a trade barrier. All key departments' servers can only use autonomous chips and autonomous operating systems.

In fact, now that the performance of computers is excessive, many areas do not necessarily require high performance. Prism doors make security issues outstanding. There are opportunities for marketization by using state-subsidized autonomous chips and autonomous operating systems throughout the year.

Can industrial base blessings allow Chinese cores to take off?

This time, the Ministry of Industry and Information Technology has engaged in a strong industrial base for 2015, and it is difficult to say how much the relationship between itself and Prism Gate is related. However, it mentions the basic components, and the high-performance CPU is definitely the key support object.

In fact, the country has been investing in Loongson for several years. However, the difference is that the security awareness of some sensitive departments behind the Prism Gate has been raised, and the international giants such as IBM and HP are all undermined. The greater market gap allows Loongson to enter, and Godson hopes to survive on the market. .

On March 30th, China launched the 17th Beidou satellite. The central processor of the satellite is "Godson."

The replacement of "Loongson" satellites is just a beginning. The number of servers required by China's security-sensitive departments is huge. If they are all replaced by Loongson 3B servers, then their business benefits can support Godson recruiting first-rate talent and developing new levels of Godson 4 and Godson. 5, but the software will gradually become popular with the popularity. Only in this way can we compete positively with international giants.

Of course, the ideal is only ideal. The replacement of servers in sensitive departments and the migration of data are not simple. This process will take place slowly and will not be immediate. No matter if it is Godson or Divine, it still has a long way to go before it completes the marketization cycle. It may take 10 years or more.

China's "core" is difficult to get on the road, and we bless it.