Aries MM-700's 3G baseband core uses Qualcomm's, for Liu Ruoyi, it is obviously a core link that is subject to others, if the scale is small in this area in the future, then of course everyone, hello, hello, everyone, there will be no bottleneck.

But once the MIPS mobile processor is scaled, it is obvious that this key baseband will become the "seven inches" key point that is stuck by others!

According to Liu Ruoyi's logic, it doesn't matter if the components and technologies that can be replaced are someone else's.

But the core irreplaceable technology must not always let others get stuck, and it is best to do it yourself if you can do it yourself.

The GPU used in the mobile processor, the source of IMG is a company with many partners, and now MIPS has been authorized after intervening in the field of mobile processors, and has also invested almost 15% of IMG's shares, becoming a strategic major shareholder of IMG.

Also major shareholders of IMG are industry giants such as intel, SS company, and moto. With such a company with diversified equity, Liu Ruoyi is not worried about being stuck in the GPU area.

However, the baseband is different, and this link is also a place where Qualcomm communication patents are quite concentrated, and it is difficult to find a replacement manufacturer if there is a problem in the future.

With this idea, Liu Ruoyi's technical team naturally went to find a solution.

There are not many companies that can do 3G baseband, but it just so happens that MTK on the island has this strength and has a cross-licensing agreement with Qualcomm. Liu Ruoyi's technology introduction team quickly set their sights on MTK.

In terms of the scale of MTK itself, the market value of NT$100 billion has just exceeded NT$100 billion at the end of the year, and the volume is not large.

It's easy to do things if you have money, MTK's equity is seriously diversified, and Liu Ruoyi first secretly absorbed chips in the secondary market through marketization.

After collecting almost 20% of the total through its various institutions, one of them raised its name and said that it would make further acquisitions and was ready to take a controlling stake in MTK.

This move has obviously attracted some attention from the industry on the island, after all, MTK is now in a unique position in this area, and it is not an ordinary company.

However, Liu Ruoyi is not an ordinary person now, and after several rounds of behind-the-scenes interest games, MTK's management is powerless to prevent the acquisition of Liu Ruoyi's offshore company.

Considering the sensitivity, on the surface, Liu Ruoyi did not use his own name to buy directly this time, but several seemingly unrelated offshore investment company puppets acted in the name of capital investment.

Therefore, there is no reason for the authorities on the island to prevent the approval of the change of shares. In this way, after Liu Ruoyi's offshore company acquired nearly 30% of the shares, the original shareholders lost control, because the shares in the secondary market of Liu Ruoyi's other offshore institutions have exceeded 50%.

If it is a meeting that votes to re-elect the board of directors, Liu Ruoyi's people can oust all the current board of directors.

Fortunately, Liu Ruoyi did not plan to remove the management, but sent a puppet under West Coast Capital to temporarily dominate the board of directors.

At the operational level, the other executive agencies and departments are now kept unchanged, and only through this puppet at MTK to promote the conclusion of cooperation agreements with MIPS.

In the future, a new processor based on MM-700 will be developed for MTK's 3G processor part, and of course the baseband part will be replaced by MTK's own 3G baseband.

Because Liu Ruoyi, the shareholder of the offshore company, has a clean background, and there is no connection with Liu Ruoyi, the industry has speculated about it, but no matter how many rumors there are, it is not clear how much there is a connection with MIPS.

This has made Qualcomm and other industry predators not pay much attention.

MTK is only in 2G through the copycat machine a big hit, 3G era of most of the key patents firmly in the hands of Qualcomm, even if MTK chip shipment does not need to pay directly to Qualcomm patent fees, but the development of MTK 3G chip mobile phone manufacturers have to pay this patent fee when selling mobile phones.

This money can't run anyway. The famous "Qualcomm tax" is not in vain.

From this point of view, the MTK chip is nothing more than a porter of Qualcomm's tax.

However, as long as there is this privilege of 3G authorization without MTK, Liu Ruoyi is already satisfied.

The offshore company here has just taken over the board of directors of MTK, and soon MTK launched a wave of new financing, ready to raise $2 billion to invest in research and development, this time Liu Ruoyi in addition to allowing MTK to continue research and evolution on the fledgling 3G baseband, the new financing funds are heavily invested in the research and development of 4G technology.

Although the application of 3G has just begun, it is necessary to take precautions in the future 4G patent war, if there is no accumulation of its own, Liu Ruoyi believes that once MTK develops to the extent that it can threaten Qualcomm, it will inevitably be coerced by the other party in patents.

What's more, 4G is the moment of the greatest real benefit, the future 4G baseband market space is much larger than the short-lived 3G, MTK has invested heavily in this area to do research and development accumulation, and the future of the 4G voice is obviously more powerful.

With this new financing, coupled with the strong support of Liu Ruoyi and his family in terms of development in the future, even if it cannot form a direct threat to Qualcomm, at least the future development space is stronger than the previous time and space.

The technological competition in this world is like this, the top of the technological food chain of the war industry swallows the biggest piece of fat in the development of the market, and there is some soup and water left for a kind of follower to eat, but that's all.

If it is grown enough to threaten the other party, then there is no warmth to speak of, and the space for development must be stifled.

Whether it is the brutal price war launched by SS in the memory war, or the patent war launched by TMS against Zhongxin International, all of them are extremely dangerous life and death battles.

Qimonda that fell in the memory war is not the first, nor is it the second, Liu Ruoyi's rise of SS Company in the last time and space is also stepping on the accumulated bones of the Oriental people to ascend to the throne of memory!

It's just that in this time and space, it encountered a hard stubble, and Liu Ruoyi was born, which is more powerful than SS Company, and the means are more intense.

In a shorter time, the greater production capacity directly crushed the last group of people in the market to insist.

Based on the already stable Dream Factory, the DRAM market that was supposed to recover in the second half of 2009 has not yet seen the opportunity to recover.

Although from the market demand side, there is already more demand.

But Liu Ruoyi was too ruthless, and forcibly increased the production capacity harvested from the island to more than 90%, which laid the foundation for the supply of ordinary DRAM market to skyrocket.

Elpida on the Oriental side has begun to complain.

The market share, which was originally the third largest in the industry, is shrinking day by day.

With unprecedented production capacity, Qimonda rushed its market share from the fourth place in the industry to the second position in the industry, surpassing Hynix and Elpida, the two previous market share leaders.

Most of the new market share is squeezed out of these two companies.

In less than a quarter of time to restore the rated capacity of the relocated production line, Qimonda has stirred up the rivers and lakes of memory.

Although Qimonda doesn't make any money in this market now, Liu Ruoyi can afford to lose.

In a bloody war, in addition to comparing the physique of both sides, it is also necessary to look at the ability of blood transfusion!

Now there are people in Qimonda who are transfusing blood while watching him shop with his competitors.

What's more, Qimonda still has its own profits in high-end display memory, and it can actually resist the past in the short term.

However, Elpida is not good, and has now taken the lead in asking the parent company for help, asking for at least $500 million in bailout funds, which is the loss data it estimates for the next six to eight quarters.

If the parent company promises to give this money, Elpida still has hope to hold out for another 2 years.

Otherwise, two more quarters from now on would cause the company to lose valuable cash flow and enter a pre-bankruptcy state.

On the other hand, Hynix has been complaining, but it has found a new backer, and it is rumored that SK has intended to eat Hynix.

In this memory market, the Koreans will not let SS Company dominate the world independently, and they now want to unite to stabilize the industrial advantage that they have won so hard.

Although most people now don't believe that they will kill Qimonda after the change of ownership, they can choose to kill the memory business of the Orientals and Yankees, making the current five-power battle in the memory industry evolve into the Romance of the Three Kingdoms!

The Koreans, who are familiar with Chinese historical allusions, are still full of confidence in the new situation.

Even if Qimonda can rush up to become Wei in the short term, SS Company and Hynix can also suppress Qimonda's offensive.

With the demand for nearly one million wafers of memory every month, Qimonda's family can't eat it!

With his current production capacity, the production capacity of 500,000 memory wafers is remarkable, and now everyone will see how to integrate the last loser in the market.

Although SS and Hynix were also suffering, they never had the idea of giving up, but began to silently grit their teeth and accumulate strength, preparing to deal with the new situation after the old pattern was broken.

Southwest.

Jianchuan Computer's 32-channel SPARC T2 prototype has been saved, which is basically designed according to the original architecture, but after using the low-power DDR2 memory of Qimonda 65nm process process, with better heat dissipation treatment, Li Shijie adjusted the SPARC T2 operating frequency to run at 1.8GHz.

At this frequency, the two contradictory factors of stability and performance are taken into account, and at the same time, the power consumption of the whole machine is about the same as that of the old server before, and even the power supply part can continue to share the original model.

It can be said that this design has considered many factors, and the balance of cost, stability, and performance has been almost brought into full play.

The first 24 hours of full-load operation test was stress-free, and the heat dissipation, memory occupation and CPU occupation of the machine were running at more than 90% of the high level, and the system remained normal and not stuck, showing the strong ability of the chip.

The only drawback is that the gigabit network with built-in processor is not yet able to be used by Jianchuan computer, and now it needs to use the gigabit network communication built into the motherboard, so that the communication ability has not been able to play its greatest effect.

Because the SPARC T2 processor already has 2 Gigabit Ethernet controllers built-in, it is fully capable of leading out for better network connectivity.

It's just that Li Shijie has limited manpower, not to mention that there is also a shortage of talents in this aspect of the network, so he still chooses the mature gigabit network module on the market to integrate it on the motherboard.

Even with such a little fly in the ointment, the results of the preliminary test were enough to amaze Song Jianjun, and he couldn't even wait to take this prototype away.

For him, Jianchuan Computer originally supported the data center, and if there was a problem, it would be repaired, anyway, the old machine before was not used like this.

Of course, Li Shijie understands the reason why the old guy is in a hurry to get this machine, a certain big guy still has less than a year to finalize, and a large amount of test data needs to be verified and improved and optimized.

Now is the time when the task is tight, and having this new guy join will at least increase their ability in this area by more than 5 times!

It's not an easy thing to do.

Li Shijie took the test indicators and studied them carefully, confirming that the probability of major problems on this server was relatively small, so he agreed to the old man's request.

Anyway, the assembly of the second and third sets is almost completed, and the first batch of motherboards will be completed in small batches of 50 servers, and then after the test is completed, it can be delivered to the Pengcheng SPARC open lab for compatibility verification testing and performance testing.

Only after getting the certificate of Pengcheng SPARC Open Lab can Li Shijie dare to get it to the financial sector to apply.

Otherwise, he would not have had a chance to enter the door of the application of the financial industry!

Gangnam.

A 16-socket SPARC T2 server was also built after testing the performance metrics.

The chief designer fell into deep thought.

He was a little skeptical about the necessity of his project.

For the SPARC T2, which is now open source, far exceeds the performance of the project chip in his hand, Zhao Mengguo, the chief design engineer, feels that he seems to have gone a little wrong.

After all, the starting point of my project is the processor architecture design of the last century. Although it is said that it is very adaptable and scalable in terms of architecture.

But after all, the source is a decade behind.

It's a pity that this complete open source of SPARC T2 came a little late.

After thinking for a long time, Zhao Mengguo sighed, his face still cloudy.

It wasn't until his number one general, Li Xiaohua, came to the door and called him a few times that Zhao Mengguo regained his senses.

"Xiaohua, what did you say just now?"

Zhao Mengguo asked a little confused.

"Mr. Zhao, this is the news I just received, the Southwest Institute has made a 32-way server, which is the performance test index data. ”

Li Xiaohua's face is not very good-looking, as one of the main designers of the project, he saw his biological "son" being completely crushed by a wild child.

It's not a good mood for anyone.

Zhao Mengguo grabbed the few pages handed over by Li Xiaohua, and his face was even more pale after reading it.

He hammered his head with his hand. pondered for a while.

"Xiaohua, what do you think about this?"

Zhao Mengguo raised his head and stared at his number one favorite general, obviously understanding how the other party was feeling now.

"Mr. Zhao, I'm afraid it's difficult for our hard steel to catch up with this CPU, I dare to say that the performance of the 32-way server of the Southwest Institute this time was announced, and a lot of orders belonged to them. Li Xiaohua said bitterly.

When it comes to SPARC servers, they're at least a few years behind their opponents.

"Moreover, I heard that their 64-socket servers are still in the development process, and the progress is still relatively smooth, and they can basically switch the application of the Solis system seamlessly, which has too much impact on us. ”

Not only is the impact great, but there is a risk of breaking the roots.

If the market for minicomputers and high-end servers is completely cut off by the other side, they will only have the extremely narrow and dangerous road of supercomputers in the future.

However, the space of this market is calculated in single digits, even if the unit price is high, how meaningful it is to make it.

What's more, Zhao Mengguo knows that if the distributed design is adopted, the design potential of the SPARC T2 supercomputer will not be weaker than their chips, and the current SPARC T2 chip performance is at least five years ahead of them!

In this time, a new generation of supercomputers was designed.

This time, the number of orders for the domestic SPARC Open Alliance is unusually large, and Zhao Mengguo feels a headache when he thinks about where the extra processor will be used.

Obviously, the friends who rely on the technical resources of the capital will not let go of such a good opportunity.

This time, this batch of processors is basically completely produced in accordance with domestic demand, and the performance is good, and the production is smooth, which can completely make people see new hope.

"Xiaohua, do you think we have a chance to join forces with the Southwest Institute?" Zhao Mengguo suddenly asked.

Li Xiaohua was stunned!