In addition to MIPS's Asia-Pacific R&D center, Sun Microsystems will also establish its own Asia-Pacific R&D center in Xiangjiang, and will set up research institutes in Beijing and Pengcheng respectively, gradually moving the R&D system to a lower-cost region.

Among them, the Asia-Pacific R&D Center is responsible for the research and development of semiconductor chips in Asia, relying on the hardware development environment that is gradually forming in Xiangjiang, to develop new models of chips for Sun Microsystems' next-generation servers.

The Beijing R&D Center is responsible for software research and development, and Sun Microsystems is the company with the longest software product line in the world, from operating systems to applications.

Pengcheng Research Institute provides technical support for SPARC's open and ecological commercial promotion, including Sparc software and hardware.

To put it simply, since Sun Microsystems open-sourced the SPARC hardware design in 2005 and launched the last generation of SPARC T2 processors in 2007, it has not continued to evolve SPARC T2, but this R&D team is still there, and Liu Ruoyi will certainly not leave them idle.

In order to squeeze the last bit of oil and water out of SPARC, Liu Ruoyi wanted them to try if they could continue to let more alliance members promote the further development of SPARC and extend the life cycle of the SPARC ecosystem through the open source model.

Pengcheng Research Institute is the main position leading the progress of these two open source projects, and Liu Ruoyi's goal in the future is to continue to evolve the SPARC system ecology through this open source, and strive to build a SPARC cloud IaaS in the future.

Sun Microsystems' goal in the future is to be a cloud service provider, and compatibility with the original Solaris applications is one of the aspects, which is to fulfill the long-term operation and maintenance commitment to the original SPARC customers, and on the other hand, the needs of the existing customers to migrate Solaris applications to the cloud.

As for Heungjiang's Asia-Pacific Research Center and the hardware R&D team in the United States, the company will develop a new generation of low-cost, high-performance, low-power server processors based on MIPS64.

In the future, Linux-based cloud services will be carried out on this MIPS64+ Linux platform, and if it can be successful, Sun Microsystem will usher in its second spring.

The new server based on the MIPS64 kernel will be significantly ahead of the X86 service in terms of power consumption, and the sun microsystem now has the famous database MySQL, web server, and even the office application staroffice, it can be said that Microsoft has something Sun Microsystem has, and even more...

If all of this becomes cloud services, then this is a huge market opportunity for cloud services.

Although Liu Ruoyi doesn't know much about technology, Amazon's cloud service and Microsoft's later cloud service are one of their pillars in the future!

The role of MIPS, only here, can we really see its great potential and role.

Only a giant with the industry's leading software and hardware R&D capabilities can take on the important task of MIPS64's evolution to high-performance and low-power server processors.

Placing this important task in the Asia-Pacific R&D center in Xiangjiang, supplemented by the American team, naturally provides some convenience for the general-purpose processor of Wu Zuhua's team.

As the main R&D team of the national team, Liu Ruoyi knows that there are some things that she has reservations, although Wu Zuhua did not make it clear why she must promote the development of general-purpose processors.

But how smart Liu Ruoyi is, it is rare to be confused by pretending not to know some things, so she will leave Wu Zuhua with R&D funds for general-purpose processors.

Even in the future, she has no opinion on all the money earned by this company to fill this hole.

Some things don't mean you have to make money to make money.

For the safety of some special applications, Liu Ruoyi has nothing to lose some economic benefits.

This is also another consideration for her to buy MIPS, after all, after the purchase, there will be more flexibility for the company's R&D layout and patent processing.

X86 has been short-circuited, there is no hope in the short term, and don't want to catch up in the long term. There are really more supporting applications, and it has to be RISC's MIPS and the previous open source SPARC, but SPARC open source is a bit late, and everyone hasn't caught up, plus the complexity of the system and the future prospects should be worse than MIPS.

At the very least, MIPS32 can be useful in lightweight embedded applications, and MIPS64 can also be useful in servers and super systems.

The commonality between the two is also relatively large, which is equivalent to taking into account the application of both major fields.

For the vast majority of aerospace, MIPS32 is the best choice, and for a more advanced command system, the early warning radar system MIPS64 is the heavy responsibility.

Wu Zuhua's team can understand the logical design of every hardware, and naturally there will be no danger that X86 may be hacked at any time.

Liu Ruoyi didn't bother to think about these things. As long as they come out of the broad road instructed by Liu Ruoyi, they will get a lot of profits from the application market in the future to subsidize the R&D investment in special market applications.

Wu Zuhua's team was quickly restructured, and according to Liu Ruoyi's investment, a group of people came out to set up a Gaosen Semiconductor Technology.

Now that he has money in hand, and Liu Ruoyi's left hand is right handed, MIPS has licensed all its various cores to Gaosen Semiconductor. Of course, the price is still settled according to the regular commercial price, and it is all the money invested by Liu Ruoyi anyway.

After Wu Zuhua's Gaosen Semiconductor got the MIPS32 74K core, he carefully evaluated and understood this core, and confirmed that it was very suitable for transformation into a high-end multi-core media processor. In this core, it can basically solve the floating point computing power needs of high-definition 1080P and Dolby surround sound processing.

The low-end video processing is sufficient to use MIPS32 24K, which supports video encoding and decoding needs below 720P, which is suitable for the needs of the current so-called high-definition wired digital set-top box.

Moreover, this core is very flexible, and the 130nm process process can produce 625MHz high-performance embedded processors to meet the needs of 720P encoding and decoding. If you want better performance, you can break through the 1GHz higher performance by replacing it with the 90nm process, and even under the 40nm process, some customers have launched the 1.4GHz ultra-high frequency M24K to complete H.264 and 1080P decoding without any problem.

For plug-in devices such as set-top boxes, of course, faster and better processing speeds can be obtained with higher frequency processors, but for cameras with a tight power supply, the smaller the power consumption, the better, and the lower the cost, the better. So at this time, the 625MHz M24K processor is of great significance.

Wu Zuhua did what he said, divided the research and development into three major parts to implement, video decoding, M74K and M24K high and low combination research and development at the same time, to provide video decoding customers with a single chip solution, replacing TI and other original expensive solutions, the cost is at least less than one-third of competitors.

M24K is relatively simple, while also reducing the cost of chip production, targeting the IPTV, HDTV terminal player and digital TV markets should be very competitive.

This piece of research and development is relatively difficult, and there has been no video decoding here, but as the first strong army in the system, Wu Zuhua also has a way, and after having money in his hand, he will directly recruit a large team to jointly tackle the problems of video decoding chips in brother institutes.

The research funds of Wu Zuhua's Gaosen Semiconductor are out, and the two sides share the technical achievements in the future, but the intellectual property rights must belong to the funder Gaosen Semiconductor.

At least give the brother institutes a research and development conditions that are not short of funds, and in the future, after making scientific and technological achievements, various selections will be of great benefit, and it will also facilitate them to apply for more financial support in the future.

But Wu Zuhua has greater benefits, a stronger R&D team, more skilled engineers, and the feasibility of directly converting the other party's manpower into his own manpower because of these projects in the future.

Wu Zuhua is still very proficient in this little abacus, and he has promised to spend money, but he is just a small head. In the future, the income after opening the market is the big head.

Wu Zuhua attaches more importance to the general-purpose CPU, got the latest MIPS64 core model, they found that the gap is really a bit big, although the things made behind closed doors can be used, but compared to other people's things, whether it is reliability, stability or performance and power consumption, it is a crushing gap.

Of course, this thing is very natural, no matter how bad MIPS is, it is also the result of decades of evolution, and there is a period of time when the shipment volume is still the largest in the world, and it is very normal to achieve this kind of accumulation.

If Wu Zuhua and them work behind closed doors for a few years, it will be abnormal to catch up with others!

In view of this, Wu Zuhua decisively overturned the original 3A design, prepared to digest and absorb it from scratch on the basis of the latest kernel design, reconstructed it after eating it thoroughly, added some new things that he explored in the process of research and development, and cut out some things that he didn't need to use, so that it was more stable and reliable in structure.

This can not only greatly save R&D time and costs, but also reduce R&D risks a lot. The so-called autonomous control, now that the hardware source code has been mastered, this is basically autonomous control.

Compared with the original 4-core 1GHz design, the new core starts with 4-core, with 8-core and 12-core high-performance options, and the technology roadmap for the next few years is to 64 cores!

Moreover, the main frequency of the single core is also directly from the 65nm process process of 1.5GHz, and if it is changed to the 45nm process, it can reach 2GHz!

In terms of supported memory, DDR2 and DDR3 can be selected to support, in order to obtain a higher performance CPU, Wu Zuhua did not hesitate to choose the 4-core 45nm process technology and DDR3 memory as the technical benchmark of the new 3A.

The first is that Zhongxin International's dragon factory has this condition, and the second is that this achievement will obviously have an explosive breakthrough impact, which will have a strategic impact on the development of this project in the future.

As a leader in the R&D system, Wu Zuhua is actually under a lot of pressure, because his competitors are also trying to catch up, although they are one step ahead as the main force of the national team, but due to different mechanisms, the financial guarantee is much worse than that of friends.

So the opponent has preemptively come up with a 1.4GHz product, the original 3A plan without Liu Ruoyi, even if the tape-out is realized, it will lag behind the opponent's single-core technical parameters, because of this, the opponent will take out 4 cores ahead of the core, and create higher computing power with more cores.

Although there are 8 cores and 12 cores to choose from, they have studied 4 cores for so long and have a greater grasp, not to mention that the single-core frequency parameters have surpassed their opponents, and they also have an overwhelming advantage in memory throughput.

Choosing 4 cores is more sure that it will be possible to achieve tape-out within this year, and if it is in front of the National Day, it is of great significance.

In order to achieve this goal, the new 3A team is already preparing for closed development.

As the leader of Gaosen Semiconductor and the leader of the CPU R&D national team, Wu Zuhua's pressure can be imagined.

This is another major reason why he is very quick to provide preferential conditions and R&D funds to recruit the video decoding research team.

Manpower is really starting to be short, and Wu Zuhua will not be sloppy if he agrees to the market indicators specified by Liu Ruo.

Otherwise, if there are enough people, he will definitely eat alone.

It's not easy to get funding in the physique now, and no one will let go easily. Finally, now that he has plenty of money, Wu Zuhua also has to keep the grain and grass in his hands from being wasted.

It was not until Gaosen Semiconductor was able to make a profit that it dared to breathe a sigh of relief.

In this sense, Wu Zuhua is really grateful to Liu Ruoyi for pointing out a possible way to make money for them.

It's just that there are still some channels for information exchange between these two friends, and the original design of the high-performance single-core processor has been obtained here, and the friendly side has also ruthlessly readjusted the research and development plan.

Originally, it was 8 cores, but considering that the main frequency of the single core can't compete here, then double the core and go to 16 cores!

With 4 times the number of cores, the floating-point processing performance of the stacked processor can reach 141 GFLOPs DP! According to the intelligence, the floating-point processing performance of the processor pushed out by the new 3A is up to 128 GFLOPs DP!

Although it looks like the plan is going to be a year behind, it doesn't matter, the performance indicators are better than the victory!

There is a secret rivalry between the capital and Jiangnan, which is difficult for outsiders to understand. Liu Ruoyi didn't even think that after she inserted a lever like this, she had already exerted some subtle changes on the CPU competition in the mainland like butterfly wings.

With a variety of missions, Wu Zuhua's team is moving forward with a heavy load.

The rearrangement of Sun Microsystem and MIPS R&D did not directly affect the cake of the original employees, and everything seemed to be uneventful.

Xiangjiang is not a very uncomfortable place for many Americans, so two Asia-Pacific R&D centers have landed here, and many people will even offer to work here for a period of time, so that they can change the environment.

For Tony Tsai, the new chairman of the board, the R&D strategic direction provided by the CEOs of both companies, the CEOs of both companies chose to execute.

In their view, this is nothing, it will neither affect their authority, nor will it cause harm to the company, the Asia-Pacific region will be one of the main markets in the future, a large number of companies have already set up branches there, and Sun Microsystems originally had R&D institutions in the mainland, but it was reduced after the decline.

Now returning to the Asia-Pacific region is also a familiar road for Schwartz.