Unconsciously, 2011 has passed, and for Liu Ruoyi's various companies, 2011 has been a year of rapid progress.

The birth of AMS has set off a bloody storm in the storage industry and launched a strong challenge to SS's position as the king of memory.

In terms of semiconductor foundry, after Zhongxin International integrated the production capacity of Elpida and Spansion, which had just been acquired, the production capacity has reached a new level.

More importantly, the 28nm process technology and the 25nm process technology have been successfully developed and mass-produced, which suddenly promoted Qimonda's Saxony factory, Dragon factory and Elpida Toyo factory to the 2Xnm process process era at the same time.

In this way, from the aspect of process technology, Zhongxin International's foundry business suddenly shortened the technical generation gap with its competitor TMS, and finally began to catch up with TMS from the process process.

Obviously this is a very landmark change.

Now these high-end process technologies have begun to mass-produce high-end performance of 1GB single capacity GDDR5 and 4GB single capacity DDR3-1866, for this batch of memory with AMS brand logo, soon passed a series of market tests and evaluations, with the speed of light transmitted by modern networks, AMS's latest video memory AMS GDDR5 tightly put the crown of the king of performance on the head, using AMS Graphics cards with GDDR5 memory are significantly more powerful than those with SS company's GDDR5 memory.

As a result, orders for high-end computer graphics card memory and PC memory poured in, and the Saxony, Dragon and Toyo factories all started to conquer the high-end memory market for AMS with a strong processing capacity of 150,000 wafers.

At the same time, it also began to contribute huge profits to Zhongxin International Xiangjiang Company, although the yield rate of the current 2Xnm process technology is just over 50%, but the foundry price is high enough to make up for the cost gap of the yield.

Moreover, every month, the factory's high-end process production line is constantly optimized and adjusted, and the yield rate is gradually rising at a rate of several percent, and it will rise to at least 75% or more in two quarters at most.

Similarly, DreamWorks in the mainland has now completely solved the yield problem of ordinary memory production, with a yield rate of more than 95 percent.

With the world's largest computer production base and demand, the mainland's memory demand is also the highest in the world, which makes DreamWorks at home take advantage of the home field. Began to squeeze the market share of SS and Hynix, which also have foundries in the mainland.

With the outbreak of production capacity and began to form a price advantage, AMS spared no expense in the ordinary DDR3-1333 and DDR3-1666 mainstream memory market, quickly broke through the 40% dividing line of market share, and successfully pulled SS company from the throne.

Despite the stubborn resistance, SS is now holding on to the 35% lifeline to fight back, and has begun to increase its high-end process production capacity to counter AMS's superiority in high-end memory.

But the other two competitors in the memory market have now begun to become embellishments, and Hynix and Micron have begun the countdown to death in this competition with a combined market share of less than 25%.

If there is no other direction and a way out, it is not surprising that Hynix, which has just changed hands, will continue to fight against AMS through the will of the country and will be incorporated into SS Company.

And magnesium, it depends on whether Sakamoto thinks he has the value of acquisition.

The CEO of Micron, who is familiar with him, has begun to contact Sakamoto to discuss the feasibility of cooperation between the two companies, and for Micron, what he is striving for now is nothing more than a good price, in this regard, although President Sakamoto is good friends with each other, but he will not exchange the interests of AMS.

With Micron's current dispensable market share, Sakamoto believes that the only value is the patents in Micron's hands, especially in mobile storage.

With such an outcome, no one could have imagined that in just three years, the storage market would undergo such huge changes in just three years.

For veterans of the industry who are familiar with memory, it is simply a snap of a finger!

For TMS, a competitor, the past 2011 was also tearful.

Now the total production capacity of the two sides is basically the same, due to the backwardness of the high-end process production capacity, the final foundry market share and revenue have also been surpassed by Zhongxin International!

After all, high-end processes create high value.

In 2011, Zhongxin International topped the semiconductor wafer foundry list with a total foundry sales revenue of 11 billion US dollars and a market sales share of 36.9%!

TMS ranked second with $10.7 billion in sales revenue and a slight gap of one percentage point in market sales share. Of course, in terms of the number of wafers foundry, TMS still stands at the top of the semiconductor processing list, and TMS still has an advantage in terms of total production capacity.

In this way, after investing more than $50 billion in a large amount, Liu Ruoyi's semiconductor foundry business finally achieved a financial accounting profit of about $5 billion, although the return on investment seems to be far away, and the capacity expansion of the new generation of process technology is more expensive.

But in this OEM world war, Liu Ruoyi has seen the dawn of victory and will never give up halfway.

In 2012, Zhongxin International's annual capacity expansion and R&D budget of up to 10 billion US dollars allowed Liang Jisong, who was finally qualified to surface after the difficult non-compete period, was so high-spirited that he ignored the angry eyes of his former employer at TMS.

The protracted intellectual property war launched by TMS continues to be staged in a tragic way every year in which both sides burn hundreds of millions of dollars of legal budgets, and TMS cannot get the exact evidence of the details of Zhongxin International's production line, and at the same time Zhongxin International is unwilling to give in, and no one can predict how long this patent war will last before it comes to fruition.

This made a group of intellectual property lawyers in the Yankees completely excited, and the top experts in intellectual property chose to join the two growing groups of lawyers to continue the battle.

After Liu Ruoyi blocked the channel of Zhongxin International to admit defeat with a huge team of lawyers, Zhang Rujin of this time and space did not care about this lawsuit, and the high budget for fighting the lawsuit alone made Zhongxin International not afraid of legal challenges at all, so that Mr. Zhang Rujin could devote himself to the operation of Zhongxin International.

Surpassing competitors is only the beginning of a new journey, and firmly controlling this market is the goal that Zhang Rujin is now pursuing.

Process evolution and capacity substitution and updating are still the two most important things for Zhongxin International.

Further expansion of production capacity is now a lower priority, and new production capacity can be naturally increased by replacing the renewal mode with process technology.

Obviously, this battle of the king of semiconductor foundries will need to be entangled for many years, and it will not be possible to see the final victory or defeat or reach a mutual compromise until one party is unable to support it.

Similarly, for Li Shijie, Zhao Mengguo and Wu Zuhua, 2011 was also a new year for them to take important steps.

The 12-core 2GHz new generation MIPS64 architecture processor GS-3B has been supported by the 28nm process technology of Dragon Factory, and the power consumption is still in the low-power range of less than 45w, but it has achieved a computing power performance index of more than 80GFlops, which further greatly improves the application value of GS-3B.

Although it is still a little far from the current mainstream processors, such as Intel's Core i7-2600, which is 112.4GFlops per second, it also lacks multimedia computing instruction support such as SSE, so it is naturally a little powerless in the game field.

But after adjusting his thinking, Wu Zuhua's thinking has changed, and he will not fight uselessly against the deep-rooted X86 at the level of ordinary PC.

Therefore, Wu Zuhua's task is to polish the product as much as possible to make it more stable and reliable, and on the other hand, to increase the special functions of industrial applications.

For example, the new generation of GS-3B, due to time problems, still retains the use of DDR2 memory, so that the focus on multi-core expansion and reliability will come up with the product in such a short time.

Of course, in the middle of the dragon factory for many times to carry out tape-out tests, it cost tens of millions of tape-out costs is also indispensable, relying on the actual tape-out test problems, than simulation is much more sufficient, so that the typical money-burning method of exchanging costs for progress is to rely on the rolling profits brought by digital core to support.

With more than 2.5 billion digital core sales in the past year, Gaosen Semiconductor's annual profit has reached the level of 900 million, and the new second-generation digital chip DM-200 series has supported the latest video encoding and decoding methods.

The most important thing is to concentrate demodulation and other chips on one chip, which greatly reduces the difficulty of digital TV production and development, and makes digital chips the main chip in domestic digital TVs.

Tens of millions of shipments have also hit the cost of a single chip of digital chips to a lower level, and Gaosen Semiconductor, which has deep pockets, is now capable of opening a third R&D front.

Wu Zuhua cooperated with the R&D team of a university of science and technology, and the miniaturized super server developed by GS-3B can already be concentrated in a single cabinet, and the performance is close to that of a new SPARC server with 32 circuits.

In line with the dual-socket single-board design method adopted by Li Shijie, at least the GS-3B board is more compact and low-power, and a dual-socket computing unit board with 16GB DDR2 memory consumes less than 100W, and only needs to use a heatsink for heat dissipation.

This heat dissipation method has too many advantages for subsequent designs.

This new server is almost overwhelming for Shijie Li's 32-socket SPARC T2 server in terms of size and power consumption at the same performance.

If both sides use Linux, Li Shijie's 32-channel SPARC T2 server may not be able to compete with this set of servers, which is now codenamed KD-100.

Thankfully, Zhao Mengguo's side didn't lag behind, and the 1.8GHz 16-core SPARC T2plus was successfully taped out and put into testing, and the new processor with the addition of cores and DDR3 memory controller has also seen a significant performance improvement.

In this way, when the new processor enters mass production and commercial use, Li Shijie can still open the performance gap between the SPARC T2PLUS server and the KD-100 server of Gaosen Semiconductor's MIPS architecture.

With more than double the processor cores and cables and a replacement for faster DDR3, the 16-core, 64-thread SPARC T2plus is at least twice as fast.

For a 32-socket, 16-core SPARC T2plus server, this performance increase will be orders of magnitude!

The use of compatible pin design makes the application of the 16-core SPARC T2plus easier, which is why the Zhao Mengguo model still follows the T2 as the main model.

And Li Shijie is not idle, two large cabinets of streamlined supercomputers have completed prototype production and preliminary testing, using 256 processors of the Shenwei-1000 supercomputer, the performance of the supercomputer across the 10 trillion supercomputer entry level.

However, considering the concise design and power consumption of 100 million operations, the Shenwei-1000 supercomputer has killed the current world record of 100 million calculations with equal power consumption!

And the system can still run the same Solaris system that supports 64-way servers!

This is the goal of Li Shijie's dominance in the Solaris ecological chain.

As for the general-purpose Linux server, now the system is in full swing in the process of transplanting and modifying, and the cooperation with Hongqi has made Li Shijie full of confidence in the Linux server market.

The two newly rising giants of domestic servers will carry their own exclusive secret weapons and begin to rudely break into the market of X86 and PowerPC servers.

For the now mainstream server manufacturers in the mainland who are immersed in X86 and PowerPC, they have some news rumors about this.

However, I can't imagine what kind of impact the market that cannot be subverted in foreign countries will have on the mainland server market after the matching of the Linux system is completed.

...

During the same period, MIPS and Sun Microsystems completed R&D and channel integration, Sun Microsystems was renamed Sun Mobile Computing SMC, and the first generation of low-power server chip M64-1620 based on MIPS64 is ready to enter mass production, this 16-core 2Ghz server is integrated into a 16GB DDR3 controller, with a floating-point computing capacity of more than 160GFlops, and the power consumption is less than 35W!

Based on this processor, Sun Mobile Computing has developed a variety of products such as 4-, 8-, and 16-way servers. The main goal is to target those data centers that offer cloud services.

The new servers are less than one-third of the power consumption of X86 servers with the same performance, and in terms of price, Sun Mobile Computing's current strategy is to follow the price of X86 servers with the same performance, offering 15 to 25 percent discounts to large customers.

This has allowed Sun Mobile Computing, a SPARC server supplier whose server price has always been much higher than that of X86 servers, to finally find its new position, and if the data center cloud service market is opened, it is obviously a new situation for Sun Mobile Computing.

At the same time, the research and development of M64-2420 has entered the preliminary design verification stage, this 24-core 2Ghz highest frequency of the new processor, in addition to 8 more cores, in addition to the DDR3 memory support capacity expanded to 32GB, for the current processor, this has kept up with the mainstream level.

At the same time, the new feature of designing the main frequency from 1.2GHz to 2GHz also allows the M64-2420 to control the power consumption very well after the performance is improved.

According to the design index, the maximum power consumption is 40W, but under the controllable mechanism of main frequency conversion, the power consumption of M64-2420 may be reduced to less than 15W for application scenarios with small business volume or unbalanced application computing power!

If this power consumption index is realized, in principle, M64-2420 is qualified for mobile computing and entering the notebook market, of course, the notebook market is special, the user-friendliness of the Linux system is not high in terms of operating system, and the gap is too big compared to the windows system, and it is impossible to apply directly in the PC, and consumers can't accept it.

Obviously, for datacenters that consume a lot of energy, M64-2420-based servers will be a very good choice.